Bias Control

Evidence quality

Highlighted claims

• The main Cox model treated NSAID exposure as time-dependent to account for exposure changes and avoid immortal time bias. — Association between exposure to non-steroidal anti-inflammatory drugs in pregnancy and miscarriage risk: a French nationwide retrospective cohort study
• The main analysis used a 3-day lag after NSAID dispensing, with miscarriages during the lag counted as unexposed. — Association between exposure to non-steroidal anti-inflammatory drugs in pregnancy and miscarriage risk: a French nationwide retrospective cohort study
• The study addressed immortal time bias and protopathic bias as methodological problems in earlier research. — Association between exposure to non-steroidal anti-inflammatory drugs in pregnancy and miscarriage risk: a French nationwide retrospective cohort study
• Longer lag periods reduced estimated hazard ratios, but the association remained statistically significant for every lag tested. — Association between exposure to non-steroidal anti-inflammatory drugs in pregnancy and miscarriage risk: a French nationwide retrospective cohort study
• The lag-period pattern supports the importance of controlling protopathic bias, though lagging exposure could underestimate immediate NSAID effects. — Association between exposure to non-steroidal anti-inflammatory drugs in pregnancy and miscarriage risk: a French nationwide retrospective cohort study