Assessing concordance between patient preferences in advance care plans and in-hospital care.

Objective The aims of this study were to assess: (1) concordance between patient preferences stated in advance care plans (ACPs) and hospital care over the subsequent 12 months; (2) change in preferences over time; (3) justifications for discordant care; and (4) effects of ACP completion on hospital utilisation. Methods A retrospective study was conducted of 198 patients with an ACP form registered with an ACP registry and tagged with a hospital unique record number. Data collected from ACP forms and hospital records comprised ACP completion and revisions, care preferences, patient characteristics and hospital care. Instances of care discordant with preferences were analysed as Type A (no ascertainable justification) and Type B (direct patient request or appropriate clinical indications). In a survivor subset, hospital utilisation was compared before and after ACP completion. Results Mean (± s.d.) patient age was 79.5±11.8 years. Patients had a mean (± s.d.) of 5.5±2.5 comorbidities and 90 (46.4%) died within the 12 months after ACP completion. Most ACPs (130; 65.5%) were completed during index hospitalisation and 13 (6.5%) underwent revision, on average, 6.8 months later, all related to rescinding request for cardiopulmonary resuscitation. Hospital care was fully concordant for 154 (77.8%) patients, with 39 (22.2%) receiving 60 instances of discordant care (15 (25%) Type A, 45 (75%) Type B), mostly related to surgical procedures (20; 33%) and intravenous fluids or antibiotics (26; 43%). Patients receiving discordant care had higher mortality (77% vs 45%; P<0.001) and more rapid response team activations (34% vs 13%; P=0.001) at 12 months than patients with concordant care. Among the 108 confirmed survivors at 12 months after ACP completion, emergency department presentations and hospital admissions per patient had decreased by ≥50% (P<0.001) and hospital days had decreased by 25% (P=0.042) compared with the 12 months before ACP completion. Conclusion Most patients completing an ACP received hospital care fully concordant with their stated preferences, with few revising their preferences over time. Discordant care mostly related to justified supportive treatments or surgical procedures. Among survivors, ACP completion was associated with decreased use of hospital care. What is known about the topic? ACPs that list patient preferences and care goals relieve family and patient distress and uncertainty regarding future care decisions as death approaches, decrease unwanted medical interventions and hospitalisations, and are associated with more patients dying at home. However, uncertainty surrounds the extent to which in-patient care provided to patients' concords with preferences stated in ACPs, which preferences are most adhered to, and whether preferences change over time, warranting revision of ACPs. What does this paper add? This retrospective study of 198 patients completing an ACP, of whom almost half died within the following 12 months, showed that more than 75% received hospital care fully concordant with their stated preferences and, for decedents, most died at their preferred place of death. Relatively few patients changed their documented preferences over time, and all changes were for less use of cardiopulmonary resuscitation (CPR). Instances of discordant care mostly related to the administration of supportive treatments or surgical procedures and most were justified on the basis of patient request or appropriate clinical indications. Among 108 survivors, the number of emergency department presentations and hospital admissions per patient at 12 months after ACP completion was half those seen in the 12 months before ACP completion, whereas hospital days per patient decreased by 25%. What are the implications for practitioners? Encouraging patients with progressive chronic disease to complete an ACP reduces their risk of receiving care they do not want, reflected in decreased use of hospital care. Preferences stated in ACPs are mostly stable over time and, if changed, tend to become more conservative in terms of CPR. Conversely, preferences stated in the ACP do not, as circumstances change, bind patients or clinicians to withholding care that relieves symptoms or prevents major morbidity in the short term.