Primary Care Consultation for Patients With Palliative Care Needs

Study aim:

The investigators aim to develop, apply, and evaluate a mix-method intervention involving training in PC and a new consultation model in primary care for patients with PC needs.

Primary objective:

To determine the impact of the intervention on the physical symptom burden for patients.

Secondary objectives:

To determine the impact of the intervention on:

1. psychological symptoms burden.
2. communication/practical issues.
3. some determinants of health care service use (number of consultations for acute disease both in the health care unit and emergency department, number of hospitalizations and referrals to hospital health care services).

Study design:

Before-after study

Study procedures:

GPs working in the centre health care region in Portugal will be asked to collaborate with the project.

GPs will receive a training program on PC built based on our previous study with GPs where the investigators identified the role of GPs in the PC setting and their training needs in PC. They will also receive training on how to apply the new consultation model.

Each GP will be asked to recruit patients with PC needs from their own list of patients using a protocol developed for that purpose. Each patient recruited will receive a medical consultation every 3 weeks for a period of 12 weeks. In both the first and last medical consultations, each GP will ask patients to complete the Integrated Palliative Outcome Scale (IPOS) patient version.

The study will integrate a pilot evaluation involving the first two patients of each GP, so that the consultation model can be improved according to the collected feedback.

The investigators intend to assess whether our mix-method intervention can improve the physical symptom burden of patients with PC needs after a 12-week intervention

Sample size:

Sample size was calculated for the main outcome (physical symptom burden) measured using the IPOS, considering the two time points (baseline and 12 weeks after). The effect of the intervention will be measured by the difference between the two means tested using t-test for paired samples, A sample size of 53 patients is estimated, considering a power of 80%, a type I error of 5%, and a medium effect size of 0.513, and accommodating a 50% loss of patients at follow-up (UCSF Clinical \& Translational Science Institute - Sample Size Calculators, available at http://www.sample-size.net). Given that the prevalence of patients with PC needs in primary care is 8% to 14%3,4 and that in Portugal each GP has, on average, 1600 to 1900 patients in their list, it is estimated that each GP will have around 128 patients with PC needs in their list. Considering the involvement of 8-10 GPs in the study, the planned sample size is feasible. Included GPs will be asked to recruit patients until the planned sample size is reached (around 6 patients per GP).

Data collection:

The following variables will be collected by the main researcher about the GPs participating in the study: age, sex, years of clinical practice, type of primary care unit in which they work, environment (rural or urban) and previous training in PC.

The following variables will be collected by each GP about the patients participating in the study using a paper form: age, sex, education, marital status, rural or urban environment, main diagnosis, and comorbidities and IPOS patient version scores. GPs will also collect, by searching patients' medical records: number of consultations for acute disease both in the health care unit and emergency department, number of hospitalizations and referrals to hospital health care services referring to 12-week pre intervention and 12-week post intervention. A 12-week timeframe for evaluating resource use before and after the intervention was chosen to assess how these variables behaved over a period equal to that of the intervention. Each form will be pseudo-anonymized by assigning a code, so that only each GP will be able to identify which patient each form belongs to.

Statistical analysis:

Data analysis will be performed using SPSS v. 27®. Categorical variables will be described by frequencies (absolutes and relatives). Normally distributed continuous variables will be summarized by mean and standard deviation. Ordinal and non-normally distributed continuous variables will be represented by median and interquartile interval. The normality of the distributions will be assessed by observation of the respective QQ-plots.

To compare the difference of the two moments (baseline and 12 weeks after) a t-test for paired samples or a Wilcoxon test will be used. Statistical significance will be set at p\<0.05.