«Citation: Dalkin, Sonia (2014) The Realist Evaluation of a Palliative Integrated Care Pathway in Primary Care: What Works, For Whom and in What ...»
To summarise, the Hospice at Home service was relatively well known, although there was one palliative care patient and one bereaved relative who hadn’t been informed about the service. Experiences of the service were all positive, including night sitters and transport services. Reasons for not using the service were mainly due to a lack of need which was often related to a strong support system and personal preferences (reasoning), but also may have included a lack of information about it in the cases of those who were not informed about it.
Commissioned services are being used variably by patients. The palliative care unit should be offered as early as possible in a patient’s palliative care journey and not used solely as a respite or unplanned option near to the end-of-life. Patients need to be aware of their options in order to have extended choices. However, it must be acknowledged that home deaths are preferred to hospice deaths (Gomes, Calanzani et al. 2011) which is why uptake may be limited in this rural locality. The Hospice at Home service was evaluated positively by those who had used it and those who did not expressed that they had not had a need to, due to strong support networks.
ICP commissioned services should be offered to all patients yet may be of most use to those with weak support networks who will not have the capacity to stay at home, thus it is essential health care professionals identify this vulnerable population early to avoid a hospital death. However, in order for services to be utilised, patients, relatives and carers must have knowledge of them, which in some cases they did not. To conclude, in the context of increased palliative care services, the palliative care unit and Hospice at Home service provided increased support for palliative care patients and their families (resource), however uptake of this support was based on patients’ and family member’s preferences and needs (reasoning).
Context: Increased need for palliative care services Additional palliative care services were required in order for patients’ preferences to be met. An alternative outlet for hospice care (palliative care unit) and increased palliative care services in the community (Hospice at Home service, complementary therapy charity) were required. The provision of the palliative care unit and Hospice at Home service (resources) met this need and allowed the mechanism to ‘fire’ for those who required additional support.
Figure 5: CMOC2, commissioned services in the ICP The locality was in need of enhanced services for palliative care (context). The provision of commissioned services such as the palliative care unit and the Hospice at Home Service have resulted in increased available support (resource) for palliative care patients and their family members. Whether patients and family members use the services is dependent on personal circumstances (reasoning), with a perceived strong support network often stated as a reason to decline the use of services (reasoning). This has resulted in variable uptake of the palliative care unit and the Hospice at Home service in the locality. CMOC2 is displayed diagrammatically in Figure 5.
CMOC3 – Continuous quality improvement and intervention use Outcome: Increased use of Integrated Care Pathway interventions Comparisons of intervention use over time indicate whether the ICP’s components are being delivered more regularly and thus whether they are becoming standard practice. Clinical audits undertaken before and after the inception of quality improvements are an effective way of assessing whether the attempts are working (Sullivan and Garland 2010). Increased use of the ICP interventions over time means that they are being used for more palliative care patients, thus more appropriately. It must be noted that from 2009 to 2010 the population who received the ICP increased by 60,000 to 78,000 due to two additional practices joining the locality group and implementing the ICP. This means that these practices would not have had the same level of input in 2010 as the practices which joined in 2009;
therefore they were not as experienced in the ICP interventions and may not have achieved the outcomes they could have if they had joined in 2009. Additionally, values for 2007, 2008 and 2009 do not include the two additional practices.
However, the calculations controls for these missing values. The interventions that have been focused on are: palliative care registration, preference discussions, the locality advance care plan, anticipatory medication and LCP use.
Palliative Care Registration A one-way repeated measures ANOVA was conducted to compare numbers of palliative care registrations from 2008 to 2012, using Death Audit data. Death Audit and MIQUEST data both provide values for palliative care registrations; Death Audit data was used as it provided an additional 2 years of data. Mauchly’s Test of Sphericity was significant (p.05), meaning that sphericity was not assumed and Greenhouse-Geisser values are reported. There was a significant effect of time on palliative care registrations (F(2.74, 30.17) = 9.93, p.001, p = 0.47). Using the guidelines proposed by Cohen (1988) for the p (0.01 = small, 0.06 = moderate, 0.14 = large effect), this result suggests a large effect size. The means and standard deviations for palliative care registrations from 2007 to 2012 are presented in Table 4, the means show an overall increase in palliative care registrations from 2008 to
2012. However there is a small decrease in the mean from 2010 to 2011. This is depicted in Figure 6.
Table 4: Descriptive statistics for palliative care registrations conducted in the GP practices signed up to the ICP over time, using Death Audit data.
Figure 6: Mean number of palliative care registrations between 2007 and 2012, using Death Audit data.
Preference Discussions A one-way repeated measures ANOVA was conducted to compare numbers of preference discussions that were documented on the data system in 2009/10, 2010/11 and 2011/12. The means and standard deviations are presented in Table 5;
the means show an increase in preference discussions recorded on the data system over time. Mauchly’s Test of Sphericity was not significant so sphericity was assumed. There was a significant effect of time on preference discussions recorded on the data system (F(2, 22) = 15.95, p.001, p = 0.59). Using the guidelines proposed by Cohen (1988) for the p (0.01 = small, 0.06 = moderate, 0.14 = large effect), this result suggests a large effect size. The increases over time in preference discussions recorded on the data system are identified in Figure 7.
Table 5: Descriptive statistics for preference discussions recorded on the data system conducted in the GP practices signed up to the ICP over time, using MIQUEST data.
Figure 7: Mean number of preference discussions between 2009/10 and 2011/12, using MIQUEST data.
The number of practices involved in the analysis (N) is twelve instead of fourteen as SPSS accounts for there being no data for two of the practices in 2009/10 as they joined the ICP at a later date. These missing values are not regarded as an issue as the reason they are missing is not related to the data or intervention.
Locality advance care plan A one-way repeated measures ANOVA was conducted to compare numbers of locality advance care plans completed in 2009/10, 2010/11 and 2011/12. There was no significant effect of time on locality advance care plans completed (F(2, 22) = 0.21, p.05, p = 0.2). This means that the number of locality advance care plans being carried out since the introduction of the ICP has not significantly increased.
Anticipatory medication In order to assess whether the use of anticipatory medication has increased over time a one-way repeated measures ANOVA was conducted. There was no significant effect for time on anticipatory medication used (F(2, 22) = 2.67, p.05, p = 0.2). Therefore, despite Figure 8 displaying that anticipatory medication is increasing, this increase is not statistically significant.
Figure 8: Anticipatory medication prescriptions from 2009/10 to 2011/12.
Liverpool Care Pathway for the Dying Patient A one-way repeated measures ANOVA was conducted to compare numbers of patients who died with the LCP. The means and standard deviations are presented in Table 6; the means show an increase in LCP use over time. Mauchly’s Test of Sphericity was not significant so sphericity was assumed. There was a significant effect for time on LCP use (F(2, 22) = 3.67, p.05, p = 0.25). Using the guidelines proposed by Cohen (1988) for the p (0.01 = small, 0.06 = moderate, 0.14 = large effect), this result suggests a large effect size. Therefore it can be concluded that the GP practices are significantly increasing their use of the LCP since the introduction of the ICP, this increase is shown in Figure 9. The means and standard deviations indicate a large increase from 2009/10 to 2010/11.
Table 6: Descriptive statistics for LCP use in the GP practices signed up to the ICP over time, using MIQUEST data.
Figure 9: Mean number of LCP use between 2009/10 and 2011/12, using MIQUEST data.
Summary of ICP intervention analysis The statistical analysis of GP practice data suggests that the use of some of the ICP interventions is increasing over time: palliative care registrations, preference discussions and LCP. All of these interventions indicated a significant increase in use from implementation of the ICP to 2012. Furthermore, Cohen’s (1988) effect sizes indicated that all of the significant increases in intervention use were large.
However, neither anticipatory medication nor locality advance care plans showed significant increases over time. In terms of locality advance care plans, this nonsignificant increase might be due to health care professionals’ difficulties with the documentation (discussed later in chapter 6, p.199).
Mechanism: Continuous Quality Improvement In the ICP CQI takes the form of PCQVs which utilise GP practice data (Death Audit and MIQUEST) to feedback information on performance to practices. It is a resource that the ICP provides, as CQI requires leadership (from the ICP’s founder) and financial resources (through ICP commissioning streams) to educate and enable health care professionals to evaluate themselves. It involves not only identifying issues with the ICP and problems in practice but also gives recognition of effort (reward systems) in both verbal and written forms (McLaughlin 1994). These clinical governance visits consist of the founder of the ICP visiting the GP practices MDT meeting to: feedback data, reflect on the practice’s performance, discuss how the practice staff would manage a fictional palliative care patient from diagnosis to death, talk about difficult issues experienced when using the ICP, and identify and provide potential solutions for problems specific to that individual practice. Thirteen of the fourteen GP practices involved in the ICP and this study have received one PCQV. The PCQV were evaluated positively by both a community matron and a GP, who felt that they received constructive feedback tailored to the individual practice (resource) which helped them to improve as a team.
Community Matron 2 (FG1): “(I’ve) been involved with two because we cover two practices, and they’ve both been very different and they’ve been constructive, very honest, but constructive and actions have happened as a consequence.” GP3 (FG1): “Yeah, (inaudible) brought the team together to discuss it with (ICP founder) and we did get very helpful feedback which we took on board” GP2 suggested that the PCQV resulted in practices engaging with the ICP more (reasoning), due to the information that they were provided with about their performance.
GP2 (FG2): “And I wondered if under input, and again it’s a more general comment, whether we should be putting something about audit, feedback, education process into this, and I was thinking specifically about (ICP founder’s) visits to practice, because I know that a lot of the practices, you know, have engaged a lot more, further, after one of (founder’s) visits because of the information that was fed back to them.” Community Matron 2 and GP2 then built on the above quotes by explaining how even just the prospect of the PCQV resulted in them improving their performance (reasoning).