This study showed that the prevalence of chronic pain was 5.8% in the HHC elderly. Nearly half of the HHC with chronic pain belonged to the 75–84 years age group. Malignancy predominated as the cause of chronic pain and more common in the 75–84 years group than other age subgroups. Female participants were more likely to have osteoporosis as the cause of chronic pain than male participants. Although they received HHC services, 76.7% of the participants were hospitalized and 52.8% of the participants visited the ED within 1 year after discharge from the hospital. The followed-up mortality was high, 64.1% within 1 year, 79.9% within 2 years, and 84.3% within 3 years, without difference between two sexes and among three age subgroups. Acetaminophen was the most commonly used analgesic and morphine was the most commonly used opioid. There was a trend of decreased use of opioids with the age.

Contrary to common belief, the prevalence of 5.8% of chronic pain in the HHC elderly is surprising low. The possible explanation is the different criteria among studies. A convenience sample of community-dwelling elderly in Taiwan reported that the prevalence of pain was 50% [17]. Another study in Taiwan reported that the prevalence of pain in nursing home elderly was 65.3% [18]. However, both of the aforementioned studies utilized definitions of pain based on the self-reports of patients, which differs from the criteria for chronic pain in this study, since the strict criteria of “using of analgesics for at least 3 months” was also used. This additional criterion may have contributed to the low prevalence rate.

In this study, malignancy was the most common cause of chronic pain in the HHC elderly. The incidence and prevalence of most malignancies increase with the age [19]. Previous studies reported that 25-40% of geriatric patients with malignancy living at nursing home had daily pain [7,20]. However, 26% of these patients with daily pain received no analgesics and patients older than 85 years were more likely to receive no analgesia, which indicates that malignancy related pain may be undertreated in the geriatric population [20]. Regularly given oral analgesics are the most common treatment for the malignancy related pain in the elderly [21]. Because the elderly are at increased risk of developing toxicity from NSAIDs, opioids should be considered when the elderly have contraindications or uncontrolled pain after NSAIDs use [21]. The elderly need lower dose of opioids, therefore, careful titration based on individual response including the side effects can ensure the appropriate response to clinical demand [21]. In addition to analgesics, adjuvant medications including antidepressants, antiepileptics, corticosteroids and bisphosphonates may help in the treatment of certain types of chronic pain [21].

Females had more musculoskeletal pain and osteoporosis than males [22-24]. The prevalence of hip osteoporosis was 2% for men compared with 16% for women [24]. In women, the prevalence of osteoporosis rapidly increased (tripled) at the age of 70 years. In men, the prevalence of osteoporosis increased at the age of 80 years and the rate doubled [24]. In addition to the higher prevalence of osteoporosis, women may have higher pain sensitivity due to the more central sensitization [25].

The hospitalization and ED visit rate within 1 year after discharge were 76.7% and 52.8% in our study. It is difficult to compare previous data with this study because there is a great variation of medical care, insurance, culture, and local policy among different research. A study based on Medicare data reported that 20% of the beneficiaries (most are elderly) are rehospitalized within 30 days and 34% are rehospitalized within 90 days [26]. Another cross-national study reported that HHC elderly had a 25.3% of hospitalization and 15.7% of ED visit rates within 90 days after discharge [27]. The goals of HHC are to assist the patients to remain at home, avoid hospitalization or admission to long-term care institutions, and improve the patients’ function to give them greater independence [28]. A previous study revealed that an acute exacerbation of chronic disease is the most common reason for an unplanned hospitalization or ED visit, which can be prevented through knowledge of risk factors, provider communication, and careful monitoring [29]. Other risk factors are polypharmacy [30], the length of HHC episode [31], worsening primary or secondary diagnosis or the development of a new problem [32], falls [30], and advancing age [30]. Strategies recommended for preventing hospitalization or ED visit are patient/caregiver education, 24-hour on-call nursing coverage, telehealth, management support, case management, front loading visits, medication management, fall prevention program, and positive physician and hospital relationships [33].

The followed-up mortality in the HHC elderly with chronic pain was high in this study, however, there is also no comparable data in the literature. A study in Spain reported that 28.9% of HHC elderly died during the one-year follow-up period [34]. The independent predictors for morality were male sex, comorbidity, degree of pressure ulcers, and having received home care rehabilitation [34]. Another study about nursing home residents in Iceland reported that 28.8% of the residents died within a year, 43.4% within two years, and 53.1% of the residents died within 3 years [35]. The possible explanation for the higher mortality found in our study is that chronic pain and its underlying disease may increase mortality.

This study has the major strengths of possessing a nationwide design and presenting delineation of an unclear issue. However, there are also some limitations. First, we used “analgesics use for at least 3 months” as the surrogate criteria of chronic pain, which may be too strict and therefore result in an underestimation of the prevalence of chronic pain. Second, there may exist multiple causes of chronic pain, and the primary cause may be difficult to determine. Third, detailed information related to morality, including activities of daily living, life style, body-mass index, and socioeconomic status are not available in the NHIRD. Fourth, despite being a nationwide study, the results of this study may not applicable to other nations due to the differences in insurance, medical resources, race, and culture. Further studies are warranted to validate these result.