Patient and family engagement in patient care and research in Canadian intensive care units: a national survey

Participant identification

We aimed to develop an exhaustive list of Canadian ICUs and PICUs, including academic and community hospitals. Two individuals (M. P.-A., M. V.) developed lists of ICU leads across Canada using a multimodal strategy (email, internet searches, and telephone with local, regional, and provincial colleagues in critical care). When identifying ICU leads, we sought individuals with a significant administrative/leadership role in the ICU, who would be able to speak to current clinical and research practices. In some instances, these self-reported individuals were physicians; in other cases, they were administrators from varied backgrounds.

Survey development

We used a systematic approach to questionnaire design and reporting.14,15,16 Research ethics approval was sought, but waived, by the Hamilton Integrated Research Ethics Board (Hamilton, Ontario, Canada).

Item generation, reduction, and formatting

Five investigators (K. B., S. O., E. M., S. D., O. S.) generated questions in two domains of patient and family engagement (patient care and research), in addition to demographic data depicting characteristics of hospitals, ICUs, and research personnel. The same investigators reduced items within domains to retain relevant items and reduce respondent burden.14 A research assistant formatted the questionnaire in both national languages (English and French) prior to administration. We included two gift cards of CAD 2 for a national coffee franchise (one for each of the patient care and research sections) as an incentive for survey completion with each questionnaire administered.

Questionnaire pilot and sensibility testing

We pilot tested the questionnaire with five intensivists (department heads or clinicians) to identify poorly worded or redundant questions and responses, and to assess questionnaire flow, salience, and acceptability.14 Subsequently, we assessed the clinical sensibility (comprehensiveness, clarity, and face validity) using a clinical sensibility tool customized for this questionnaire and the time required to complete the questionnaire (20 min) with four intensivists.14 We sought to achieve diverse responses (French, English, men, women, academic, community, pediatric, and adult intensivists) among pilot and clinical sensibility testers. The final questionnaire was formatted in English and French and included 11 questions plus nine demographic questions with nominal, ordinal, and interval response formats (available as Electronic Supplementary Material [ESM] eAppendix).

Questionnaire administration

We sampled all identified Canadian ICUs (convenience sample) without formal sample size estimation. Prior to questionnaire administration, we sent prenotification e-mails (on 26 September 2019, and again on 12 November 2019, for e-mails that bounced back). We administered the initial questionnaire to critical care site leads on 25 November 2019, by post. Three volunteers (M. D. S., S. D., W. D. S.) created survey packages including a cover letter, paper questionnaire, prepaid addressed return envelope, and gift cards. Subsequently, we administered one electronic reminder to critical care site leads with a fillable PDF version of the questionnaire identical in appearance to the paper questionnaire (separated by approximately four weeks). We requested that the critical care site leads identify one individual in each ICU who could complete the patient care section (e.g., practice leaders, clinical lead managers, physician, or nurse leaders) and a second individual who could complete the research section (e.g., research manager/coordinator, physician, or nurse leader). We acknowledged that in certain clinical settings, the same individual may be best suited to complete both sections. The ICU demographic data section could be completed by either the individual completing the domain pertaining to engagement in patient care or research or the critical care chief/lead. Participation was voluntary and questionnaire completion implied consent to participate.

Statistical analyses

Using our list of potential respondent adult and pediatric ICUs in Canada, we planned to administer 232 questionnaires and aimed for a 50% response rate.17,18 We summarized binary and continuous data using counts, percentages, and medians with interquartile ranges [IQRs], where appropriate. Surveys with missing items were excluded from the denominator for that response. We used the Mann–Whitney U test to compare nonparametric continuous or ordinal data.

In subgroup analyses, we compared whether engagement was a priority in clinical care and research (responses to Q1 and Q6, respectively), based on practice setting (academic and community hospital with a university affiliation vs community hospital without a university affiliation), type of ICU (adult vs pediatric), presence of dedicated research personnel (full- or part-time), and presence of a dedicated individual for promoting and supporting engagement. We used the rank biserial correlation test to identify factors associated with engagement as a priority in patient care and research, wherein Somers’ delta (range, -1 to +1) provides the strength of the correlation between each factor (dichotomous variable) and engagement priority (ordinal variable). For example, an independent variable that had a statistically significant delta of 0.44 on engagement in research improves the accuracy of predicting engagement in research as a priority by 44%. All analyses were conducted in IBM SPSS Statistics for Windows, Version 26.0. (IBM Corp., Armonk, NY, USA).

留言 (0)

沒有登入
gif