The PREference STudy of lanreOtide autogel (PRESTO) was a multinational, multicentre, noninterventional, simulated-use study involving nurses with experience managing patients with NETs and/or acromegaly by injecting long-acting SSAs. This study was performed at clinical or market research centres in countries where the LAN new syringe was approved prior to or during the testing period (if during, testing sessions were initiated after the approval date): Belgium, Germany, Italy, Spain, the UK and the USA. All centres allowed for the shipment of study drug to these locations and storage in refrigerators.
Nurses attended a single testing session between June 2019 and September 2019, during which they injected injection pads (Remedy Simulation Group, Perkasie, PA, USA, or 3B Scientific, Somerset, UK) with each syringe twice (randomised injection order) before reporting their preferences. Data were collected using an anonymous, self-administered, cross-sectional, web-based questionnaire on a tablet device.
Recruitment and Eligibility Criteria
Nurses from organisations and institutions where patients with acromegaly and/or NETs were routinely treated were targeted for recruitment. Identified nurses were sent an invitation email, which contained a nurse identification form to assess eligibility and qualifications. Further details of the recruitment process are available in the supplementary materials.
Eligible participants were adults aged ≥ 18 years, employed as a hospital- or community-based nurse, or nurse from a local physician’s office or nurses’ network, with ≥ 2 years’ current and/or previous experience administering long-acting SSA injections to patients with acromegaly and/or NETs (≥ 4 acromegaly and/or NETs patients per year; or ≥ 1 every 3 months). No specific exclusion criteria were defined.
All enrolled nurses signed a questionnaire participation consent form to participate in testing sessions and to complete the questionnaire.
Eligible nurses were invited to attend a 60–90-min testing session (up to seven nurses per session). Sessions were standardised across testing centres and conducted in the local language (English, French, Flemish, German, Italian or Spanish). At these sessions, clinical research associates (CRAs) provided nurses with injection-related testing resources, including leaflets with information about each syringe and a slide deck that included a study overview, and detailed instructions for testing the syringes and completing the electronic questionnaire.
Nurses were provided with materials to administer four simulated injections (two with each product: LAN 120 mg and octreotide LAR 20 or 30 mg) into two injection pads (one per product). Both syringes were provided by Ipsen, contained real products and were labelled ‘not for human use’ on the box in their local language. LAN new syringe packs consisted of a pre-filled, ready-to-use syringe (0.5 ml with an automatic safety system and a 1.2 × 20 mm needle) in a laminated pouch. Octreotide LAR syringe packs contained: one 6-ml glass vial with a rubber stopper, sealed with an aluminium flip-off seal, containing powder for suspension for injection; one 3 ml pre-filled glass syringe with a front and plunger stopper with 2 ml solvent, co-packaged in a sealed blister tray; one vial adapter; one safety injection needle. Before each session, products were retrieved from the refrigerator at least 30 min prior to the testing as per instructions for use on the label [10, 11].
The order of injections with each product was randomised using a pre-populated randomisation list with block sizes of two to ensure counter balance in the ordering of syringes throughout each session and the study. CRAs at each session were given a unique section of the randomisation list to prepopulate the session log, and nurses were assigned a randomisation number consecutively upon arrival.
After testing each syringe twice, nurses completed the web-based questionnaire using a tablet device. CRAs completed the session and randomisation number fields for each participant. Once the questionnaire was completed, data were saved on a secure central server.
Objectives and Assessments
To develop the current web-based questionnaire, a systematic literature review, detailed in the supplementary materials, was performed to identify relevant previous preference studies that utilised a questionnaire methodology with additional input from a patient-reported outcomes expert (DC) and nurses (including DA). The web-based questionnaire, comprised of a single four-level preference question, and the rating and ranking of nine pre-defined attributes, is provided in full in the supplementary materials.
The primary objective of PRESTO was to assess nurses’ preference for the LAN new syringe compared with the octreotide LAR syringe after performing injections into injection pads. This was assessed by the proportion of nurses who stated a strong/slight preference for each syringe in response to an overall preference item with four response options (slight or strong preference for the LAN new syringe or the octreotide LAR syringe).
Secondary objectives were:
To describe nurse preferences for the attributes of each syringe, evaluated through rating nine pre-defined attributes using a 5-point Likert-type response scale from 1 (not at all) to 5 (very much).
To describe the importance of syringe attributes. Based on the same nine pre-defined attributes, nurses were then asked to rank the three most important attributes (in rank order) and one least important attribute. Nurses were also asked if any attributes they considered important were not listed.
To describe the clinical characteristics of nurses (e.g., clinical experience, setting).
To describe the sociodemographics of nurses (e.g., age, gender).
The exploratory objective was to assess how syringe attributes and factors such as nurse characteristics and sociodemographics related to overall preference.
During nurse sessions, an issue was identified: a number of injections administered using the octreotide LAR syringe resulted in clogging, found to be due to the injection pad, which was consequently changed during the course of the study. Further details on this issue are described in the supplementary materials.
This study aimed to recruit 90 nurses based on a meaningful difference of approximately 15 percentage points and a preference of 57.5% for one syringe and 42.5% for the other syringe. This sample size was based on using a one-sample binomial test with a power of 81% and two-sided 5% alpha level. This sample size was determined prior to initiating recruitment. Analyses were carried out on the enrolled analysis set (all nurses), consisting of all participants who self-reported meeting all eligibility criteria, as well as by injection pad used, defined as the first pad analysis set (FPAS) and second pad analysis set (SPAS).
The primary outcome was analysed using a one-sample binomial test for proportions of nurses preferring the LAN new syringe compared with the octreotide LAR syringe. The importance of attributes for each syringe are summarised descriptively through frequency of responses, related percentages and the difference in rating between the two syringes using paired t tests. Analysis of syringe attributes was also categorised by scores of < 5 and 5 (i.e., highest rating, 5 = very much), also known as a ‘top-box’ score [12,13,14,15]. Rankings for each attribute are summarised descriptively through frequencies and percentages indicating the first, second and third most important and the least important.
Descriptive analyses were conducted for clinical characteristics and sociodemographics of nurses, including frequencies and percentages for categorical variables.
Exploratory analyses were performed using univariate logistic regression models to identify parameters associated with overall preference using a significance level of P < 0.20, presented as the odds ratio (OR; 95% confidence interval [CI]). Collinearity of these selected variables was examined through correlations, ANOVAs, chi-square or Fisher tests, as appropriate. Variables in the univariable logistic regression model with a significance level of P < 0.2 were considered for inclusion in a multivariable model.
All data were forced responses; thus, there were no missing data in variables captured. All analyses were performed using SAS® v9.3 or later (SAS Institute, Cary, NC, USA).
As this study was not a clinical trial and no patients were involved, institutional review board approvals were not required. This study followed the recommendations from the International Society for Pharmacoepidemiology, Good Pharmacoepidemiology Practice Guidelines, April 2007. The confidential nature of personal information was maintained and participating nurses provided authorisation for the use and disclosure of personal information (e.g. age, gender, clinical experience) in the study.