INTRODUCTION
Patients' participation in care seems to be important for ensuring that patients receive high quality care based on certain values, such as autonomy. Patients are supposed to be actively involved in care and treatment, even though such situations are often complex, especially in the context of end-of-life care (Singer et al., Reference Singer, Martin and Kelner1999; Steinhauser et al., Reference Steinhauser, Christakis and Clipp2000; Schroder et al., Reference Schroder, Ahlstrom and Larsson2006). In Sweden today, the fact that patients have the right and autonomy to make their own choices makes it necessary for healthcare professionals to collaborate with them (SFS, 1982, p. 763). Studies investigating patients' participation in end-of-life care are rare, because of ethical and methodological reasons such as exposing these patients for stressful research (Ingleton & Seymour, Reference Ingleton and Seymour2001; Hudson et al., Reference Hudson, Krisjanson and Quinn2005). Patients' participation in care and treatment forms a foundation for dignity, by creating feelings of confidence and comprehension, and allowing them to maintain a sense of control (Eldh et al., Reference Eldh, Ehnfors and Ekman2004).
Patient participation means a person's involvement in their own life situation. It includes making choices even when one cannot perform certain activities alone. Patients' own will, personal goals, motivations, and roles are important for providing quality of care according to World Health Organization's definition of patients' participation (World Health Organization, 2001). Participation includes the concepts of taking part, being included and/or engaged in an area of life, being accepted, or having access to the necessary resources (Cahill, Reference Cahill1998).
The interaction between healthcare professionals and a patient, concerning the type and extent of care and treatment she/he wishes, is a crucial aspect of patient participation. Patient participation is dependent upon what the healthcare professional knows about the individual's wishes, so that she/he can make it possible for the patient to collaborate in clinical decisions based on values such as autonomy. Patients must be adequately informed by healthcare professionals to make it possible for them to participate in different healthcare situations (Sahlberg-Blom et al., Reference Sahlberg-Blom, Ternestedt and Johansson2000; Bottorff et al., Reference Bottorff, Steele and Davies2000; Sainio et al., Reference Sainio, Lauri and Eriksson2001). A patient's wishes and needs are supposed to be documented in their medical and nursing records and should correspond to what type of healthcare the patient really wants, which may not be what they actually received (SFS,1993:20; SFS,1982:763)
There is a growing consensus that patients' participation in care (i.e., daily life choices as well as decisions about medical procedures) needs to be boosted even in the end-of-life care. Despite the fact that patients are enthusiastically encouraged to participate in this type of care, there is a need for a broader research field, which addresses this issue as well as decision-making about different medical treatments. It is important to acquire considerably more knowledge about the circumstances surrounding patients' participation in care at the end-of-life, since this has been shown to have a marked influence on dignity and well-being at this time of life (Chochinov, Reference Chochinov2006; Randers & Mattiasson, Reference Randers and Mattiasson2004). The aim of this study was to identify demographic and health-related variables' relation to patients' participation during the last 3 months of life, as documented in patients' records.
METHOD
Design and Study Sample
A retrospective review of death certificates, medical records, and nursing records comprise the foundation for population-related information. The population in the present study consists of 229 patients from 49 municipalities in the County of Västra Götaland in Sweden. The number of deaths in the county was 15,869 persons (7,733 males and 8,136 females), during 2001. Data were collected from all available documentation about deceased patients, and the inclusion criteria were: people who had died in the County of Västra Götaland during 2001, aged ≥18 years at the time of death, and who had received healthcare services according to the Health and Medical Service Act, during the last 3 months of their life.
Two hundred and eighty individuals were chosen as a primary sample. The first selection was made in proportion to size (PPS), i.e., proportional to the number of death incidents in 10 municipalities within the actual county. The next step, a random selection of individuals from chosen municipalities, was made in PPS. Two extra municipalities were included in the sample to ensure sufficient representation and guard against an underrepresentative sample from the county. Twenty-five individuals were excluded because medical and/or nursing records were inaccessible. Eighteen were excluded due to having died from sudden death, accidents, or suicide. Eight were excluded because no healthcare services had been employed during the last 3 months of life. The final study sample consisted of 229 individuals.
Data Collection
Before the data were collected, a tool, based on the framework by Donaldson and Field (Reference Donaldson and Field1998) and by Stewart et al. (Reference Stewart, Teno and Patrick1999), was developed. This framework emphasizes problems relating to quality as well as to healthcare indicators. One of these quality indicators focuses on patients' participation in different decisions during the end-of-life phase and was used as a foundation for selection of the data in this study. Patients' participation in everyday life, place of care, pain relief, medical examinations, and treatments were domains used in the protocol as a base for data collection. The four domains developed from a pilot study, in which all records from 20 deceased patients were examined. The protocol was first pretested and the results from these pretests led to revisions, improvements, and retesting of the items, until an inter-rater agreement reliability of >85–95% was reached. All data regarding demographic and clinical characteristics relevant to the patients' last 3 months of life were collected. The collection described occurrence of different variables in the medical and nursing records. Descriptions in the patients' records showing participation was identified from texts articulating the patients' will and wishes. The actual data were noted in the protocol after reading all of the patients' medical and nursing records as well as their death certificates. All available healthcare documentations during the patients' last 3 months were examined according to demographic and health-related variables.
Ethical Considerations
The Helsinki Declaration was followed in this study (Helsinki, 2004). The Ethical Committee, University of Gothenburg approved the project (Ö633-02).
Statistical Analysis
Two groups were compared; patients who were described as participating in care and patients whose records lack such information. These two groups are described by using frequencies. Comparisons between the groups are assessed with the Fishers exact test. Univariate analyses were performed in a first stage to select the most relevant variables in relation to the patients' participation. All variables that showed a p value <0.1 were included in the multiple linear logistic regression models and this model was used to assess which of the variables were predictive of patients participating in care. The significant main effects in the multiple linear logistic regression models were assessed using the backward stepwise method. This method starts by filling the full model, including all variables, and then removes unimportant variables one at a time until all those remaining in the model contribute significantly. Unless otherwise specified, all tests were two-tailed, and a p value of <0.05 was judged as significant. The Hosmer and Lemeshow test was used for assessing goodness-of-fit (Hosmer et al., Reference Hosmer, Hosmer and Le Cessie1997)
RESULTS
In this study the mean age among the patients was 80 years; half of them residing in private homes and the other half in residential care facilities. One hundred five (46%) of the included patients were female and 124 (54%) were male. One hundred and thirty-two (58%) of the patients were described as participating actively in their healthcare. Patients' participation as documented in their records varied from occurring only in one situation to occurring in several situations during the last 3 months of their lives. In 97 (42%) of the patients' records there were no descriptions found about the patient's own will and wishes. No significant differences were found between the two groups regarding age, gender and residence (Table 1).
Table 1. Demographic variables for the total number of deceased patients comparing the group of patients participating with the group that were not described participating. P-value shows the significant differences between the groups
at-test; bFishers exact test.
The differences between the groups in the first investigation showed that patients described as participating in end-of-life care had significantly more neoplasm, musculoskeletal disease, and other symptoms or problems (pain, sleeplessness, fatigue, feeding problems, nausea/vomiting, urinary incontinence, anxiety, depression, ankle swelling, diarrhea, and various degrees of deterioration) than the group that did not participate. They also had more hospital-based inpatient care and care in private homes. Patients not described as participating had significantly more mental disorder (dementia and cognitive impairment) and disorientation regarding time, room, or person. Table 2 shows all these variables for the total number of deceased patients, with the group of patients that participated compared with the group that did not participate.
Table 2. Health Related variables for the total number of deceased patients participating or not participating at the end-of-life
In the next investigation odds ratios and 95% confidence intervals were made and are presented to illustrate the relation between significant predicting variables and the patients' participation (Table 3). Depression is the most closely related variable for patients participating in end-of-life care followed by urinary incontinence, pain, musculoskeletal disease, fatigue, and nausea/vomiting. Disorientation to time, room or person is negatively connected to patients' chances to participate in care and treatment at the end-of-life. Fifty-nine profiles identified among the 229 patients show the patients' different degrees of probability for participating in care at the end-of-life when a combination of predictive variables occurred. The most typical profile of patient participation in end-of-life care is the one with an estimated probability of 98%, which includes: patients with depression, urinary incontinence, musculoskeletal disease, pain, fatigue, and nausea/vomiting. The lowest probability profile with an estimated probability of 4% is for disoriented patients with no other symptoms/problems and diseases described in the documentation.
Table 3. The significant relations between patients' participation in care and health related variables including 95% confident interval (N = 229)
DISCUSSION
Almost 50 % of the medical and nursing records lack any information that described patients' participation in care at the end-of-life. The study showed that these patients were significantly more likely to suffer from dementia and were disoriented to time, location, and/or person compared with the patients who were involved in some way or other in their care and treatment. It is a well-known fact that patients with cognitive dysfunction are one of the most vulnerable groups regarding participation in healthcare, resulting in inadequate information regarding patients' wishes, which may lead to insufficient palliative treatment (Nolan et al., Reference Nolan, Hughes and Narendra2005; Aminoff & Adunsky, Reference Aminoff and Adunsky2005; Mahon & Sorrell, Reference Mahon and Sorrell2008). To ensure quality of care, it is necessary to assess disoriented patients' wishes, as well as to plan carefully in order to establish special skills, knowledge, and empathy among the healthcare professionals involved (Eldh et al., Reference Eldh, Ekman and Ehnfors2006; van Thiel & van Delden, Reference van Thiel and van Delden2001). Black and Osman (Reference Black and Osman2005) point out the importance of making a distinction between patients' capacity and competence. Although involvement in care for patients with cognitive dysfunction is problematic, this does not mean that it is impossible, but further investigation is needed to find innovative solutions to this problem, if the necessity of patients' active involvement in care is ever to become a reality for everybody (Holmerova et al., Reference Holmerova, Juraskova and Kalvach2007). According to documentation in healthcare records, we found indication that the more problematic patients' situations relate to symptoms/problems, the more they are involved in care and treatments. Some of the most common symptoms/problems for patients at the end-of-life are pain, fatigue, depression, nausea/vomiting, and urinary incontinence (Nordgren & Sorensen, Reference Nordgren and Sorensen2003).
Patients with disorientation to time, location, and/or person had fewer symptoms documented, which has been shown in several other studies. This means that subsequently, patients with cognitive dysfunction are underestimated regarding symptoms/problems at the end-of-life (Hall et al., Reference Hall, Schroder and Weaver2002; Klinkenberg et al., Reference Klinkenberg, Willems and van der Wal2004; Oi-Ling et al., Reference Oi-Ling, Man-Wah and Kam-Hung2005). This could be seen as natural but the problem is that patients with cognitive impairment do not express their wishes in the same way as people who are cognitively unimpaired (Chochinov, Reference Chochinov2006). Our results indicate that the documentation lacks important information that might make these patients more involved in their care and treatments.
Inadequate notes in the patients' records could indeed point to a problem in documentation, which is recognized as a discrepancy between patients' expressions and healthcare professionals' descriptions, especially when related to patients with cognitive dysfunction at the end-of-life (Hegarty et al., Reference Hegarty, Hammond and Parish2005; Nordgren & Sorensen, Reference Nordgren and Sorensen2003). It is important to develop improved standards for documentation, and policies for healthcare professionals, in order to deal with the wishes of patients and to guarantee them quality of care at the end-of-life. Further research related to patients' participation in end-of-life care is necessary, and special attention must be focused on patients with different kinds of cognitive dysfunction.
Study Limitations
We have chosen to study participation in healthcare through retrospective reviews of medical and nursing records. Findings must be interpreted in the context of the strengths and limitations of such data sources. Studies have shown that there are discrepancies between the content of patients' healthcare records and patients' self-reported problems. Notes in documentation regarding patients' participation are short and sometimes inadequate (Nordgren & Sorensen, Reference Nordgren and Sorensen2003; Hegarty et al., Reference Hegarty, Hammond and Parish2005; Kirchhoff et al., Reference Kirchhoff, Anumandla and Foth2004). Furthermore, this specific data set represents a combination of subjective symptoms voiced by the individuals themselves, and indicators by professional caregivers.
CONCLUSIONS
This study demonstrates patients' participation in end-of-life care as it was noted in the patients' documentation. Patients with dementia and disorientation to time, location, and/or person lacked a variety of descriptions in their records about their will and wishes. Cognitively unimpaired patients' problem/symptoms and wishes were described in a extensive way.
These findings demonstrate that descriptions in the records about patient participation were not visible for patients with cognitive dysfunction, as compared with records for those patients who were cognitively unimpaired. Even if involvement in care for patients with cognitive dysfunction is problematic, it does not mean that it is impossible. Further investigation is needed to find innovative solutions to this problem. There is a need for improvement of documentation showing the patients involvement in care, and for a discussion on how to assure patients a high quality care at end-of-life even if patients' voices are not heard.
ACKNOWLEDGMENTS
We thank the Audit Department at the County Council of Västra Götaland for their financial support, the County and municipalities for helping us to realize the study, and Nibia Aires, Ph.D., Doctor of Mathematics, for her invaluable support with statistics.
