There appeared to be a lack of a systematic approach to the recording of discussions with patients or carers about end of life care issues. Even when discussions were recorded researchers auditing the notes and those interviewed identified there was substantial variability in terms of the content, depth and detail of the discussion. Lack of detail is in and of itself not necessarily a problem particularly if any decision made is clearly recorded for example a preference about where one would like to die. However, lack of detail and contextual information as to how and why a decision was arrived at and what other issues have been addressed and discussed in terms of end of life care preferences is cause for concern if an individual moves across care boundaries and away from the care team with whom they had their initial discussions, something that is likely to happen in the approach to death.
For cases without evidence of end of life care discussions in the last four weeks of life, it was generally unclear as to whether discussions had either not taken place or were simply not recorded. It may be that patients were unable to engage in discussions at this time due to their mental and/or physical condition. However, this was not clear from the case notes. In a number of cases researchers encountered end of life care related discussions that were recorded prior to the last four weeks of life. These were not counted for the audit because they fell outside of the specified sampling window, and this is a potential limitation of the study design. However, it could also be argued that discussions should occur and be recorded at regular intervals as there is evidence from this and previous studies that patients' choices and concerns are not always clear cut and are likely to change over time, especially as death becomes imminent [15, 20]. There needs to be a systematic and regular approach to reviewing any recorded wishes about end of life care already in place.
The findings presented above raise a number of concerns, not least because inconsistent or incomplete recording of patient expressed preferences means an individual's wishes may be unknown or ambiguous. If an individual subsequently loses capacity to decide for themselves, these wishes cannot reliably inform 'best interests' decisions . In addition, poorly recorded or lack of recorded views and wishes may cause problems for patients when they move between services or into the care of different professionals within the same service. Patient choices and concerns may not be passed on or can be poorly communicated. This is highly undesirable given that where expressed preferences are clearly known they are more likely to be acted upon resulting in improved outcomes for both patients and family members . The audit findings would also support this claim in that for the majority of cases with a recorded preference, the sampled services had enabled the patient to be in their preferred place of care (typically, their home/permanent residence).
It is recognised that service providers could easily be over-burdened with record-keeping and our group interview data clearly indicate that lack of time was already impacting on the nature of what was actually recorded about end of life care discussions in patients' notes. In addition, while end of life care tools appeared to be a useful prompt and support for staff initiating end of life care discussions, there was some evidence that they were regarded as a formal task to be routinely accomplished. This finding has been reported elsewhere [14, 15] and may lead to feelings of pressure and obligation rather than leaving staff with the confidence to apply their personal discretion and judgement. However, more systematic approaches to recording end of life discussions are probably required in order to avoid the situation where information is simply retained in individual memories.
The results from our group interviews would suggest that participants recognised that records of patients' last four weeks of care with respect to discussions relating to end of life care were limited and that improvements could be made. In particular, participants suggested that their record-keeping did not reveal the depth and quality of work actually undertaken, particularly in relation to conversations held with patients around preferred place of death and other aspects of end of life discussions. In addition, it was evident that there was some confusion about what it was necessary to record. If health care professionals were clearer about why information was needed and what purposes it served, it may help them in making decisions about how much detail to record and what issues to focus on.
There are a number of limitations associated with this study. It is acknowledged that audit data reflects recording of discussions rather than actual discussions. We cannot assume that all patients were told about their condition or were aware of the nature of their situation or of the focus of the services they were receiving. Records may be incomplete or erroneous representations of services delivered. However, note-keeping may be revealing about perceived priorities, at least in terms of written communication and long-term documentation. Furthermore, the temporal specificity of the audit must be stressed. For the purposes of the audit, data collection was limited to care received in the last four weeks of life as we wanted to identify whether discussions were taking place at the very end of life. It may be, for example, that there were advanced discussions for some cases that were recorded prior to the last 4 weeks of life; these discussions would not be captured in the audit. In addition, due to limitations of time and resources, the researchers sampled from records that were kept by and within the services that they had access to; they did not obtain notes from all services with which a particular case had contact in the last four weeks of life.