In recent years, the need to provide home medical care services for older adults has been recognized in several countries,^1-4) as the population of homebound older adults has been increasing. In general, older adults who receive home medical care services have difficulty accessing hospitals independently due to their decline in activities of daily living (ADL)⁵⁾ and multiple chronic diseases.⁶⁾ During the transition period from outpatient/hospital to home care, the goals of treatment and care often change in older adults, and reevaluation of the adequacy of drug prescriptions is important.

The use of potentially inappropriate medication (PIM) is prevalent worldwide,⁷⁾ and over the past decade, the importance of providing appropriate pharmacotherapy for older adults has been recognized among healthcare professionals. Screening tools such as Beers Criteria⁸⁾ and STOPP/START⁹⁾ have been introduced to avoid PIM. In Japan, guidelines for medication review in older adults were published in 2015,¹⁰⁾ and there has been increasing use of these guidelines among clinicians and pharmacists.¹¹⁾ It has been suggested that a deprescribing process is conducted by multidisciplinary teams such as physicians, nurses, and pharmacists, which requires constant follow-up. Home medical care involves monitoring and sharing information about the patient by multidisciplinary teams. In addition, the Japanese government has introduced an additional medical fee for comprehensive medical home care services¹²⁾ that promotes comprehensive medication management. Therefore, being in a home medical care setting, especially during the transition period from outpatient/hospital to home care, can be an incentive for deprescribing. However, polypharmacy and PIMs are prevalent in home-care settings.^2,13-15) Despite the recognition of the guidelines and tools in clinical practices, little is known about how medication management is conducted in home care settings. Furthermore, polypharmacy has been reported to be associated with worse physical function and increased healthcare utilization, such as among older adults^16-18); however, there has been a lack of research focusing on clinical outcomes in home care settings. Based on these backgrounds, the purposes of this study were to reveal demographic data and medication use for care-dependent older adults receiving home medical care and to evaluate whether PIM prescriptions were associated with hospitalization and death.

This study was a retrospective cohort study using data of older adults aged ≥65 years who received home medical care, using a healthcare insurance and long-term care (LTC) insurance claims database of Kure city, Japan, in April 2017. All citizens aged ≥65 years in Japan are covered by a universal healthcare coverage system, managed by local governments. The data included age, sex, recorded diagnosis on medical claims, level of LTC needs according to the LTC insurance system in Japan, and medication profile. Recorded diagnoses on medical claims were examined based on the International Classification of Diseases, 10th Revision (ICD-10) codes. Regarding LTC, people aged ≥65 years who need support for daily care are able to apply. After thorough evaluations and once they get qualified, they are classified into seven levels according to their level of disability: support level 1&2, and care level 1 to 5. Those who receive support level usually have mild disability, maintaining their independence to perform tasks needed for daily life. For people with care levels, their conditions usually require more care, as a higher number indicates a greater need for care (e.g., care level 5 requires more care than care level 4). In this study, subjects were categorized into (1) support levels 1&2, (2) care levels 1&2, and (3) care levels 3–5. It is important to note that home medical care can be provided only to those who have difficulty accessing hospitals due to their decline in ADL and their medical conditions. Physicians regularly prescribe medications at the time of the visit, which is provided once or twice a month in Japan, depending on the patient’s medical conditions. Therefore, we counted the medications that were prescribed for ≥14 days and did not include medications during hospitalizations. Pro re nata medication, medical patches, and injections were not included in this study. Compound medications were counted as two or three medications, as appropriate. PIM (≥1 PIM) was identified by STOPP-J.¹⁰⁾ The numbers of hospital admissions and deaths were identified between April 2017 and April 2019. While the reasons for hospital admissions were identified based on the recorded diagnosis, the causes of death were not identified.

A total of 2,052 participants (mean age 86.5±7.4 years, female 72.2%) were included. According to the diagnoses of chronic disease identified by the ICD-10 codes, the most common disease was hypertension (67.5%), followed by dementia (47.3%), heart failure (48.7%), and diabetes mellitus (40.3%).

The rate of LTC level was as follows: support levels (13.9%), care levels 1&2 (35.5%), care levels 3–5 (42.0%). The mean number of prescribed medications was 6.6±4.3 and the mean number of PIM prescriptions was 0.6±0.4. The prevalence of medications listed in the Anticholinergic Risk Scale (ARS) was 30.1% (Table 1).

Prescription of PIM occurred in more than half of the participants in all LTC levels—8.7% (support level 1&2), 22.6% (care level 1&2), and 26.0% (care level 3–5).

Among PIM, H2 receptor antagonists were the most common (29.0%), followed by antiplatelet agents (22.6%), magnesium oxide (19.4%), non-benzodiazepine sedatives (17.7%), and benzodiazepines (16.8%) (Table 1). The mean number of hospital admissions was 1.37±1.7. The hospitalization rate within 2 years was 64.5%. For those with hospital admission, the most common reason was heart failure (37.1%), followed by pneumonia (32.7%), urinary tract infection (16.4%) and dehydration (16.4%). Among the participants, 31.3% died between April 2017 and April 2019.

Regarding hospital admissions, number of medications (OR=1.03, 95% CI 1.01–1.06), PIM (OR=1.27, 95% CI 1.03–1.58), and care level 1&2 (OR=1.57, 95% CI 1.19–2.08) were significantly associated (Table 2). There was no significant correlation between each PIM and hospital admissions. Female sex was associated with a lower likelihood of hospitalization (OR=0.78, 95% CI 0.63–0.97). Regarding death, while age (OR=1.05, 95% CI 1.04–1.07) was associated with a higher likelihood, female sex (OR=0.52, 95% CI 0.42–0.64), care level 1&2 (OR=0.51, 95% CI 0.38–0.69), and care level 3–5 (OR=0.53, 95% CI 0.43–0.65) were associated with lower likelihood (Table 3).

This study showed the demographic data of older adults aged ≥65 years who received home medical care, and evaluated the association between hospitalization, death, and PIM prescription, using a healthcare insurance and LTC insurance claims database of Kure city, Japan. PIM was prevalent among homebound older adults. The use of H2 receptor antagonists was the most common. Number of medications, PIM, and care level 1&2 were associated with a higher likelihood of hospital admission. Age was associated with a higher likelihood of death, while female sex and higher level of LTC were associated with a lower likelihood of death.

Previous studies have reported that polypharmacy and PIM were observed among older adults receiving home medical care.^13,14,19) Although a previous study reported that the use of H2 blockers and benzodiazepines in home medical care settings decreased from 2015 to 2019,²⁰⁾ the use of these medications still remained high in our study, implying that appropriate deprescribing might not be fully established. It is worthwhile to note that the Japanese government recommends comprehensive medication management in home care services. Therefore, home care physicians are often in charge of prescribing all the medications, which enable them to evaluate each medication thoroughly, and appropriate deprescribing might have been conducted. Nevertheless, this result shows that deprescribing of some medications is often challenging. For H2 blockers, this may be appropriate, given that H2 blockers may be prescribed concomitantly with NSAIDs or aspirin to prevent upper gastrointestinal bleeding. Moreover, a previous study reported that older adults tend to have multiple symptoms, such as pain and gastrointestinal symptoms²¹⁾; thus, these medications might be prescribed for pain management as well as prevention. For benzodiazepines, deprescribing is recommended by many guidelines, such as Beers Criteria⁸⁾ and STOPP/START.⁹⁾ Despite the consensus that benzodiazepine use should be avoided, the prescription rate remained at a certain level in this study, which indicates the difficulty of managing sleep disorders. There may be some obstacles that might prevent physicians from deprescribing. Since it requires slow tapering to minimize withdrawal symptoms, physicians might feel resistance to deprescribing, given that the time to benefit from deprescribing may be longer than the patient’s remaining life expectancy. Rasmussen et al.²²⁾ reported that patients showed willingness to deprescribing of benzodiazepines, while physicians considered that patients would resist. Therefore, it is important for medical professionals to have discussions about deprescribing with patients at routine medical visits, so that patients may ask or agree to discontinue the medication. Antiplatelet agents are only beneficial for secondary cardiovascular prevention; therefore, deprescribing for primary prevention should be conducted. It is speculated that prescribing physicians might be hesitant about deprescribing, regardless of the level of prevention, because of the possibility of causing cardiovascular events, including stroke. Although patients who receive home medical care may have limited life expectancy, as one third of the patients died during the study period, a decline in ADL and lowered quality of life must be avoided; thus, the antiplatelet agents might have been continued. This finding was generally consistent with those of previous studies.^15,23,24) Therefore, utilizing the guidelines and disseminating the evidence should be strengthened for appropriate deprescribing.

Consistent with earlier research,^25,26) our study also showed that PIM was associated with hospital admissions. There was no significant correlation between each PIM and hospitalization, indicating that prescribing one PIM might not be harmful, but multiple PIMs were associated with increased adverse outcomes, which was also reported in previous studies.^27-29) A study in Thailand found that while the use of PIM was not associated with an increased risk of rehospitalization, certain PIM groups based on the 2023 AGS Beers Criteria, which included drugs that require caution, were associated with an increased risk of rehospitalization. This finding indicates that certain PIM groups may have clinical impacts; therefore, it may be critical to evaluate the outcomes of each PIM group.³⁰⁾ In addition, our study showed that a higher level of LTC was associated with a lower likelihood of death, which suggests that medical professionals should recognize that those older adults who receive home medical care might be at the stage of limited life expectancy, regardless of their level of care. When patients are in the transition period from hospital or outpatient to home care, reevaluating the goals of treatment and care is important. Shrestha et al.³¹⁾ reported that deprescribing interventions improved medication appropriateness and had potential for mortality reduction in older adults with limited life expectancy. Taken together, these results indicate that when there is a transition, careful medication review should be conducted so that unnecessary medications can be deprescribed.

This study has some limitations. First, detailed clinical information, such as the severity of the diseases, was not recorded because we used a healthcare insurance and LTC insurance claims database. Second, the diagnoses based on the claims data were not well validated. The database study is not suited to collect detailed clinical data, identifying the exact information about the medications; thus, we could not ensure whether these medications could be regarded as PIMs. Rather, they might have been appropriately prescribed to manage symptoms. Lastly, the findings of our study may not be applicable to other countries due to different clinical practices as well as healthcare systems.

In conclusion, this study revealed demographic data and medication use for care-dependent older adults receiving home medical care. PIM was prevalent in this population, and there is still room for deprescribing, suggesting that careful medication review should be conducted at the start of the transition from outpatient to home care. While there is a need for building more evidence, the guidelines should be strengthened for appropriate pharmacotherapy in this population.