Discussion
Potentially preventable medication-related ED visits and hospital admissions are unfortunately common [1-4]. From prospective observational studies, the median prevalence rate of hospital admissions associated with medication nonadherence is around 4% with almost all considered preventable [5].
In a prospective multicenter study involving twenty-one Dutch hospitals, 12,793 unplanned hospital admissions were screened with 714 (5.6%) determined to be medication-related (either from harm due to adverse effects or due to any error in the “process of [medication] prescribing, dispensing, or administering”). For those admissions categorized as preventable (332/714, 46.1%), a case-control design was used to determine potential risk factors; and in the multivariate analysis, nonadherence to medication regimen (odds ratio 2.3; 95% confidence interval 1.4-3.8), dependent living situation (3.0; 1.4-6.5), impaired cognition (11.9; 3.9-36.3), and polypharmacy (2.7; 1.6-4.4) defined as greater than or equal the use of five chronic medications, were among the statistically significant determinants of medication-related preventable admissions [3].
Returning to our case, how might the omission of levothyroxine have been prevented prior to admission ? It is possible that by decreasing the number of home medications this patient was taking (and the daughter was subsequently responsible for) and/or having a different reminder packaging systems (e.g., pre-filled blister packs from a pharmacy), may have prevented this episode of nonadherence leading to hospital admission. However, due to a scarcity of high-quality evidence and the need for more well-designed randomized controlled trials, optimal interventions to improve medication adherence in older adults prescribed multiple medications remains largely unknown [6].
How might the preadmission omission of levothyroxine have been detected earlier in this patient’s hospital encounter ? To decrease errors in the admission medication history due to nonadherence (i.e., recording that a patient is taking a particular medication preadmission, when in fact, they are not) a review of medication fill data may be helpful. This data can be obtained by directly contacting a patient’s pharmacy or pharmacies, or by having one’s institution purchase medication insurance claims data that can be viewed electronically. In our case, obtaining the fill history in addition to the caregiver interview and review of institutional medication records could have revealed this patient’s lack of a recent levothyroxine refill. However, even when armed with fill information, caution must still be exercised as even though a medication has been filled, a patient may not be taking it or may be taking it differently than prescribed.
Recently, a similar case of medication nonadherence masquerading as treatment failure was reported. Jethwa et al describe a 58-year-old male admitted to the hospital for deep vein thrombosis. Three weeks prior, the patient had been hospitalized for pulmonary emboli and started on apixaban, and although he described a 2-day period of nonadherence immediately after discharge, he reported strict compliance since that time. With his new deep vein thrombosis, treatment failure was the leading diagnosis on the differential. However, the medical team later consulted pharmacy to review the patient’s outpatient medication fill history and only at that time did the patient’s persistent nonadherence become clear. Adherence verification prevented mislabeling the patient as failing apixaban and avoided warfarin therapy and bridging with a parenteral anticoagulant [7].