Information meetings on end-of-life care for older people by the general practitioner to stimulate advance care planning: a pre-post evaluation study

A pre- post evaluation study using written questionnaires distributed and filled in at the start of the meeting (T₀), at the end of the meeting (T₁) and 6 months after the meeting (T₂) to evaluate the information meetings.

The information meetings were held by GPs from four different group practices in the provinces Flevoland and Noord-Holland, and were held from April until August 2016. The participating GPs decided to target all patients of 75 years and older registered in their practices, as older people form a relevant group for ACP and they felt it important to limit the potential number of participants per meeting. Patients were invited by their own GP by a written letter; they could enrol themselves by returning a paper slip. They could come with a companion. Sometimes this was somebody, e.g. a partner of almost 75 or with a chronic illness, who themselves were also interested in the meeting. Other times it was somebody, for example a child, who only came along to accompany a participant.

Three information meetings were held in the GP practice building, one in a church (because the GP practice was too small). The meetings lasted about 2 h, started with a powerpoint presentation on end-of-life topics and ACP by the GP, followed by time for questions. After the first meeting, the presentation was adapted to provide more information on palliative care. In the following three meetings, the GPs in that meeting used the same adapted powerpoint presentation.

When arriving at the venue, before the meeting started, the researchers handed out the questionnaires for T₀, and T₁. Attendees were asked by the researchers to fill in the first questionnaire before the meeting started (T₀). At the end of the meeting they were asked to fill in a second questionnaire (T₁). The researchers collected the T₀ and T₁ questionnaires when the participants left the meeting.

Participants were also asked to fill in a consent form if they consented to the GP sending them a last questionnaire 6 months after the meeting (T₂). If so, the GP sent the questionnaire after 6 months to the respondents home address and respondents could return the questionnaire to the researchers using a pre-addressed return envelope. With this procedure, the researchers did not have identifying data about the respondents, and GPs did not have the completed questionnaires. The questionnaires had a unique identification number to keep track of the response. Data were stored on a secured network drive of Amsterdam UMC that is only accessible to the researchers.

The questionnaire at the start of the meeting (T₀) contained questions on expectations of the meeting and reasons for coming to the meeting, questions on health, illness and quality of life, trust in the GP, previous experience with ACP, and demographics. The questionnaire at the end of the meeting (T₁) contained questions evaluating the meeting, ACP related plans after the meeting, and trust in the GP. The questionnaire 6 months after the meeting (T₂) had questions on health, illness and quality of life, trust in the GP, experience with ACP, and demographics. All questionnaires were specifically developed for this study. See the additional file for the questionnaires.

Descriptive statistics were used to summarise the characteristics of attendees. Answers to the open questions on expectations and the evaluation of the meeting were categorised. Experiences with the meeting were analysed using the Crosstabs procedure with the Pearson Chi-Square test. Pre- and post-test differences in trust were tested with a non-parametric test for paired samples (marginal homogeneity test). Because of small percentages in the categories ‘not much trust’ and ‘no trust’ these were grouped together. Pre- and post-test differences in ACP conversations and Advance Directives were analysed in McNemar-test for paired samples performed on attendees who filled in a questionnaire before and 6 months after the information meeting. In Figs. 1, 2 and 3 the questions on plans regarding talking to a physician, others and writing wishes down at T₁ were combined with whether or not an ACP conversation with the physicians or others had taken place or whether preferences were written down at T₂. This was derived from the answers on the seven ACP topics. If the answer for at least one of the seven topics was ‘I have discussed this with a physician’ than it was concluded that an ACP conversation with the physician had taken place. The same was done for talking to others and putting wishes down in writing. We did not impute missing observations. Statistical analyses were done with SPSS version 20.

In total 225 older people attended a meeting (Table 1) of which 154 (68%) filled in the questionnaire at T₀ and 145 (64%) filled in the questionnaire at T₁. After six months, 90 of the 121 people who approved of being sent another questionnaire at T₂, returned it (40%).

The majority of attendees were female (61%) and the mean age of attendees was 80 years (Table 2). Although only people aged 75 years or older were invited, some attendees were younger, for instance because the invited person came with a partner or other companion. The majority perceived their health as good or very good (84%).

More than half of attendees attended the meeting because they wanted to know more about the end-of-life (57%) or because they think about the end-of-life (52%). Most attendees expected information on practices to hasten death (euthanasia, assisted suicide, voluntarily stopping eating and drinking (VSED) (28%) and possibilities of palliative care (21%). They were asked immediately after the meeting (T₁) if the meeting matched their expectations, which was the case for a majority (63%) of participants. They could explain their answer regarding expectations in an open question. A total of 85 comments were given, which were categorised into 113 codes. Most (67%) answers were positive, for instance that the meeting was good or that they thought the information was clear. A total of 18 (9%) comments were less positive, mostly pertaining to inaudibility of the questions asked (one meeting was in a church and the acoustics were bad, in another meeting only one microphone was available which made interaction problematic) or lengthy elaborations of other attendees on personal matters during the discussion. The information on Advance Directives (21%) and the importance of talking to family, friends and care providers (15%) were mentioned most when asked about the most important thing they had heard.

The meetings were evaluated positively on the topics discussed (21% very good; 72% good), clarity of information (30% very good; 56% good), possibility to ask questions (29% very good; 68% good) and answers to questions posed (26% very good; 63% good). With regard to clarity of the information and answers to questions meeting B appeared to be evaluated more favourably, compared to meetings A, C and D (Supplementary Table A.1, Additional File 1).

Six months after the meeting attendees more often had thought about (68% vs 89%), discussed with physicians (5% vs 18%) or others (10% vs 36%) and written down (7% vs 18%) preferences regarding treatments under certain circumstances, compared to before (Table 3). Attendees more often thought about (63% vs 81%) hospitalisation and discussed this with another person (not a physician) (8% vs 29%) more often after the meeting compared to before.

Figures 1  and 3 show that attendees who plan to speak to their physicians (n = 56) or plan to put down their wishes in writing (n = 59) have done so six months later in 39% (n = 22) and 36% (n = 21) of the cases. Attendees with plans to discuss wishes with others (n = 58) have acted accordingly in 78% (n = 45) of the cases, and a further 12 out of 14 attendees who did not initially plan to do so did discuss their preferences with others (Fig. 2).

Attendees trusted the physicians more to provide good care at the end of life (p-value 0.002) and to follow end-of-life wishes (p-value < 0.001) after the meeting, compared to before (Table 4). This difference disappeared after six months. Posthoc analyses showed that attendees who had spoken to a physician after six months scored higher on trust to provide good care (p-value 0.003) and to follow end-of-life wishes (p-value 0.002) six months after the meeting, compared to attendees who had not spoken to a physician.

There are few studies on information meetings [17, 23], and we did not find previous studies of end-of-life information meetings initiated by GPs. The information meetings were held by GPs from four different group practices. Participating GPs were self-selected; they felt palliative care is important. GPs in general may be less inclined to organise meetings, and may be less capable to give the presentation and answer questions themselves. Also, we did not have information on how many people received an information letter and therefore could not study what percentage of invited people took up the invitation and attended the meeting. Attendees may have been more interested in palliative care. Those less comfortable with the topic were probably less likely to accept the invitation, resulting in a selection bias for the study and a diminished reach of the meetings. Future studies should explore strategies to help bring a broader selection of patients to the meetings.