U of T-Led BMJ Study Identifies 24 Prescribing Cascades Putting Older Adults at Risk

A population-level study of 2.3 million community-dwelling Ontario seniors, published in The BMJ on September 10, 2026, has identified 24 "potentially inappropriate prescribing cascades" (PIPCs) -- common chains where a drug's side effect gets mistaken for a new health problem and treated with a second, often unnecessary prescription. The research was led by geriatrician Paula Rochon of Sinai Health and the University of Toronto's Temerty Faculty of Medicine, working with an international team and using Ontario's ICES health administrative data.

Key takeaways

• Researchers screened 65 candidate prescribing cascades -- drawn from a 2025 international Delphi consensus of experts in geriatric medicine and clinical pharmacology -- against real prescribing data for 2,297,942 Ontario adults aged 66 and older.
• 24 cascades met all three criteria for population-level concern: the initial drug is common (used by 5% or more of seniors), the cascade itself is frequent (occurring in 1% or more of new users), and the timing pattern between the two drugs is statistically strong.
• Cardiovascular drugs most often started a cascade; central nervous system drugs did so more often in women than men.
• The authors explicitly call for pharmacists to be "more directly integrated into the prescribing process alongside physicians" to catch these patterns.

What a prescribing cascade looks like

A prescribing cascade begins when a drug's adverse effect is misread as a new medical condition rather than traced back to its source, prompting a second prescription that may not have been needed. A commonly cited example: an NSAID prescribed for pain raises blood pressure, and the resulting hypertension is treated with a new antihypertensive instead of reviewing the original pain medication. Older adults are especially exposed because they are more likely to be on multiple drugs for multiple conditions, making it harder for a patient or clinician to connect a new symptom to an existing prescription.

The three most frequent cascades identified over a one-year follow-up were iron supplement to laxative, a statin (HMG-CoA reductase inhibitor) to a pain reliever, and a cholinesterase inhibitor to a sleep aid. The strongest timing associations -- meaning the second drug most reliably followed the first in a pattern consistent with a cascade -- were corticosteroid to antipsychotic, laxative to antidiarrheal, and cholinesterase inhibitor to antiemetic.

Evidence and limitations

The study used prescription sequence symmetry analysis -- comparing how often the second drug followed the first against a background ("null effect") rate -- rather than tracking individual patient charts, so it flags statistical patterns rather than confirmed cause-and-effect. The authors are explicit that a cascade found this way is only "potentially inappropriate": some, such as adding a drug to manage a known and accepted side effect of a needed medication, can be clinically appropriate. The analysis also excluded over-the-counter drugs and cascades involving more than two drugs, used a one-year detection window that may not capture slower-developing cascades, and drew on Ontario-only data from a single, publicly funded drug benefit system, which may limit how well the specific rates generalize elsewhere in Canada.

Practice guidance for pharmacists

• During medication reviews, ask whether a current drug was started to treat a symptom that could actually be a side effect of another drug already on the profile.
• Watch specifically for the highest-incidence pairs -- iron supplement to laxative, statin to pain reliever, and cholinesterase inhibitor to sleep aid -- during MedsChecks, admissions and discharge reconciliations.
• Flag corticosteroid-to-antipsychotic and laxative-to-antidiarrheal sequences, which showed the strongest timing association in the study and may warrant a closer look even when each drug seems individually justified.
• Record why a drug was started, not only what a patient is currently taking, so a future reviewer can trace the sequence of events.
• Raise a suspected cascade with the prescriber as a discussion, rather than simply adding to it; the study's authors are calling for pharmacists to be built directly into the prescribing process, not just consulted after the fact.
• Treat a flagged cascade as a prompt for review, not automatic deprescribing -- some second prescriptions are a deliberate, appropriate way to keep a patient on a needed first drug.

What's next

The researchers say the prioritized list could be built into clinical decision-support tools that alert a prescriber or pharmacist in real time when a new prescription matches a known cascade pattern. The team, funded by the Canadian Institutes of Health Research and supported by ICES, says its next phase will focus on patient- and provider-facing education tools built around the findings.