To the Editor: One of the most distressing challenges in current geriatric practice is deciding whether to initiate artificial nutrition for elderly adults who are unable to eat. Several studies have indicated the need to improve the decision-making process for artificial nutrition.1-3 In Japan, the last decade saw increasing concern about the medical and ethical appropriateness of tube feeding in elderly adults with advanced conditions, but it appears that most medical institutions are not equipped to support the decision-making process. An explorative retrospective study on the factors affecting the decision to provide artificial nutrition in elderly adults was conducted. Fifty-nine subjects were selected from individuals admitted to St. Francis Hospital between September 2010 and February 2012. Inclusion criteria were aged 60 and older, dysphagia, and being given the decision to initiate or withhold artificial nutrition. Percutaneous endoscopic gastrostomy (PEG) was performed in 30 and enterostomy in two, total parenteral nutrition (TPN) was given in eight, nasogastric tube feeding (NGT) was administered in seven, and artificial nutrition was withheld from 12. A comparison was made between participants who did not start artificial nutrition (withholding group) and those who underwent PEG or enterostomy (tube-feeding group). Those who received TPN or NGT were excluded from the analysis because they were generally regarded as temporary treatments. Participant characteristics were compared according to sex, age, primary disease, length of hospital stay, mortality, serum albumin levels, swallowing function, physical activity, and communicative ability. The assessment methods for the last three variables have been described elsewhere.4 Analysis of the decision-making process was based on a review of participants' medical records. First, the physician's explanations about the risks and benefits of artificial nutrition were extracted from the medical records. (Analyzable data were obtained from only 23 subjects in the tube-feeding group.) Then, each explanation was typed individually onto a card. The card was then randomly presented to a speech and language therapist unrelated to the hospital, and each explanation was subsequently classified into one of two categories: positive recommendation or nonpositive recommendation. Participant characteristics of the two groups are shown in Table 1. The number of subjects in the withholding and tube-feeding groups was reduced because of lack of data for the following variables: swallowing assessment (n = 10, 15), physical activity (n = 11, 12), and communication ability (n = 11, 20). Although subjects in the withholding group were significantly older than those in the tube-feeding group (P = .04), there were no significant differences in the other participant characteristics. The distribution of primary diseases was not significantly different between the withholding and tube-feeding groups: respiratory diseases (50%, 56%), gastrointestinal diseases (33%, 16%), cardiac diseases (8%, 3%), orthopedic diseases (8%, 6%), cerebrovascular diseases (0%, 6%), and other (0%, 13%). Participants were not involved in the decision-making process because of cognitive problems, and no advance directives were reported. Families made all of the decisions with the support of physicians. Physicians provided positive recommendations more frequently for the tube-feeding group (19/23, 83%) than for the withholding group (5/12, 42%) (P = .01). Participant characteristics were not important factors in the decision-making process because there were no significant differences between the two groups except for age. Given that participant wishes were not known, the families' beliefs and physician's suggestion were considered to primarily influence the decision-making process. The analysis of the physicians' suggestions showed that they provided fewer positive recommendations for the withholding group than for the tube-feeding group. Although this might be attributable to their philosophy of patient care, an alternate point of view can be assumed in that the physicians, who after sensing reluctance from the family toward initiating artificial nutrition, were inclined to make nonpositive suggestions. Previous studies have reported that families' preferences influenced physicians in the decision-making process.5, 6 If this is the case, including a viewpoint of interactive effects between physician and family is important for improving the decision-making process. Further studies are required to create an ethically sound decision-making process that can contribute to the wellness of elderly adults who may require artificial nutrition. Conflict of Interest: The authors have no financial or any other kind of personal conflicts with this report. Author Contributions: Yoshitoshi Kuroda and Riko Kuroda were the authors of this report. Sponsor's Role: Financial support for this study was not provided.