Recurrent adverse drug reaction-related hospital admissions : prevalence, patients, predictors and presentations

The World Health Organization (WHO) defines adverse drug reactions (ADRs) as “Any response to a drug which is noxious and unintended, and which occurs at doses normally used in man for prophylaxis, diagnosis or therapy of disease, or for the modification of physiological function.” It has previously been reported that approximately 7% of hospital admissions are due to community-acquired ADRs. Older patients are at highest risk, with those aged 75 years and over exhibiting a more than four-fold increased risk of ADR-related admissions compared with patients aged 55 to 64 years. According to a recent Tasmanian study, nearly 20% of unplanned admissions to medical wards in patients aged 65 years and over were ADR-related. It can be estimated that the financial burden of ADR-related admissions in Australia exceeds AU$1 billion annually. As such, improving medication safety has become a national and international priority, with Quality Use of Medicines and Medicine Safety becoming Australia’s 10th National Health Priority Area (NHPA) in 2019.

The incidence of hospital admissions associated with ADRs is expected to increase in coming decades in parallel with ageing of the population, the increased life expectancy of comorbid patients and the escalating use of medications. The propagation of multiple ADRs in individual patients may compound this predicted escalation in ADR-related admissions because patients who have experienced an ADR appear to be at increased risk of a subsequent ADR.
To address the paucity of research that has focused on patients who are hospitalised on multiple occasions due to ADRs, the objectives of this thesis were to:

• identify risk factors for experiencing an ADR-related admission within 90 days of discharge from an acute hospital admission;
• compare the risk of an ADR-related admission for up to five years after discharge from an ADR versus non-ADR-related index admission;
• assess the proportion of ADR-related admissions that are experienced by patients with a previous ADR-related admission;
• identify the drug classes associated with repeat ADR-related admissions; and
• characterise the patients at risk of two or more ADR-related admissions within 12 months.

This retrospective cohort study was based on Tasmanian public hospitals’ Admitted Patient Care National Minimum Data Set (APC-NMDS) admission records from January 2011 until May 2017. The APC-NMDS captures clinical and administrative data elements mandated for collection by public hospitals in Australia; clinical information pertaining to each admission is coded based on the International Statistical Classification of Diseases and Related Health Problems, tenth edition, Australian Modification (ICD-10-AM). Acute admissions were classified as ADR-related based on ICD-10-AM diagnostic codes indicating that an ADR was present at the time of admission. The data set included approximately 200,000 acute admission records and 90,000 patients.

Approximately 4% of all acute hospital admissions were coded as being ADR-related. A diagnosis of cancer and an ADR-related index admission were the strongest predictors of experiencing an ADR-related admission within 90 days of discharge from an acute admission. Patients with an ADR-related index admission exhibited a more than four-fold increased risk of experiencing an ADR-related admission in their first year of follow-up, compared to patients whose index admission was non-ADR-related. A significantly increased risk persisted for at least 5 years. One-in-six ADR-related admissions occurred in patients with a previous ADR-related admission. Almost half of these admissions were associated with the same causative drug class as a previous ADR-related admission experienced by the patient during the study period. The patients most likely to experience a repeat ADR-related admission were those with an ADR-related admission attributed to antineoplastics, immunosuppressants, corticosteroids, psychotropics (excluding antidepressants), anticoagulants, and antidiabetics. There was a strong association between ADR-related admission history and the risk of subsequent ADR-related admissions; with each additional ADR-related admission patients experienced in the 365-day lead time prior to an acute admission, their risk of an ADR-related admission within 90 days of discharge from the acute admission more than doubled. Cancer, mild liver disease and younger age were independently associated with increased odds of experiencing multiple ADR-related admissions within 12 months. There was no significant association between age or comorbidity burden and the increased risk associated with cancer. In patients without cancer, a higher burden of comorbid disease significantly increased the odds of an ADR-related admission in the year after discharge from the index ADR-related admission. Patients aged 45–54 years demonstrated an almost 90% increase in their odds of readmission compared to patients aged 75 years or older, and the odds doubled in patients younger than 45 years when compared to the cohort aged 75 years or older.

The findings of this thesis should act as a catalyst for healthcare professionals across the continuum of care to better address the increased risk exhibited by patients with a prior ADR-related admission, particularly when a drug remains clinically indicated after the manifestation of an initial ADR that required hospitalisation. Initiatives that support a multidisciplinary and coordinated approach to medication management, especially in multimorbid patients and patients discharged from hospital, should be a key plank of proposed health system reforms in Australia and provide an avenue for the burden of ADR-related admissions to be addressed. Individual medication management strategies should be informed by knowledge of the patient’s ADR history and supported by evidence-based guidelines that address the patient’s comorbidity burden and treatment goals, that may change with age. Patients with cancer require particularly close monitoring in the community given the potential adverse effects of many agents used in this condition.

Alarmingly, the results of this thesis may be the tip of the iceberg since the research was based on administrative data; evidence suggests this methodology only captures between 18 and 35% of ADR-related admissions that are identified in prospective studies. Healthcare professionals may need better support to identify ADRs during routine care in order to prevent a recurrence. Future research should aim to better elucidate the root causes and preventability of ADRs that lead to recurrent hospital admissions so that these gaps in patient care can be addressed.