Study settings and population
The study was conducted as a pre- and post-intervention design from September 2020 to April 2021. The study population consisted of family members of patients who had been hospitalized in the two general medical wards of the National Taiwan University Hospital (NTUH), a large tertiary-care referral medical center in Northern Taiwan. These two medical wards were specifically dedicated to accommodating patients from the emergency care unit. Patients with any diagnosis were eligible. We included family members with: a) an age ≥ 20 years and b) adequate knowledge of the Chinese language. The study was conducted in line with the principles stated in the Declaration of Helsinki, and the study protocol was exempted from ethics review by the Research Ethics Committee of the NTUH.
The survey consisted of questions regarding demography and family members’ prior experience with a family meeting along with a 10-item questionnaire assessing the level of satisfaction. The questionnaire was aimed at rating participants’ level of satisfaction with family meetings in terms of communication, shared decision-making, understanding, and respect. The 10 items were adapted from the institutional satisfaction survey of the NTUH and modified to suit the scenario. The family members could indicate on a five-point Likert scale how they agreed or disagreed with the statement. The five-point Likert scale comprised the following responses: strongly agree, agree, neither agree nor disagree, disagree, and strongly disagree, and the corresponding scores of 100, 75, 50, 25, and 0, respectively, were assigned to the responses for statistical analysis.
The first draft of the questionnaire was created by the study team and six reviewers reviewed the relevance and wording of each item using a 4-point rating scale. The panel consisted of six experts from various academic and working backgrounds, including one MD–PhD in Health Policy and Management, one MD–PhD in Epidemiology and Preventive Medicine, one MD with expertise in multidisciplinary patient care, one RN–PhD in Psychiatric Epidemiology and Mental Health, one RN–PhD with expertise in healthcare quality and patient safety, and one RN–MS with expertise in patient referral management.
For the evaluation of the relevance, the responses on the 4-point scale were as follows: 1 = not relevant; 2 = somewhat relevant; 3 = quite relevant; and 4 = very relevant [15,16,17]. The content validity of each item was determined using the content validity index (CVI), calculated as the number of panel members providing a rating of 3 or 4 divided by the total number of the panel members [16, 17]. The clarity of the wording was judged based on the following responses: 1 = not clear; 2 = somewhat clear; 3 = quite clear; and 4 = highly clear . The face validity of the questionnaire items was determined using the face validity index (FVI), calculated as the proportion of panel members giving an item a clarity rating of 3 or 4 . The acceptable CVI and FVI at the item-level were both set at a minimum of 0.80 [16,17,18]. The CVI scores assigned by the panel members were 0.83 for questionnaire items 6 and 8, and 1.00 for the other items. The FVI scores rated by the panel members were 0.83 for items 2, 3, and 7, and 1.00 for the others. The second draft of the questionnaire was created after incorporating minor revisions in terms of language based on the panel members’ opinions.
Subsequently, the questionnaire was pilot-tested to evaluate its internal consistency and inter-rater reliability. We included 30 sessions of family meetings at this stage. For each family meeting, two family members were requested to fill the questionnaire. The questionnaire showed good internal consistency (Cronbach’s α = 0.797). We also calculated the inter-rater reliability using the intraclass correlation coefficient (ICC) and we found a strong agreement between family members (ICC = 0.763). The questionnaire was finalized after pilot-testing (Additional file 1).
The intervention consisted of two parts. The first was a website, the communication channel, containing an online blank form through which family members could submit their questions or comments on patient care. On the website, a few sentences explained the purpose of the study and encouraged family members to raise patient-associated issues. A QR code was generated to aid in easily accessing the website. The second was a written point-by-point response to the queries that family members posted on the website. The responses were drafted by the resident or nurse practitioner and finalized by the attending physician. The study procedures are narrated in detail below.
Before the intervention, family members raised their questions and issues in a verbal or written format as per their needs prior to conducting a family meeting. The attending healthcare team usually answered the questions or discussed the issues verbally in the meeting. During the post-intervention period, a pre-emptive question and answer platform was used. At least 24 h before convening the family meeting, family members were provided with the QR code to connect to the website on which they left questions and issues that they wanted addressed. The healthcare team prepared the answers to the questions or the information addressing the issues in a written format in advance. At the beginning of the family meeting, family members were provided with the documents and their pre-emptive questions and issues were clarified after introducing family members and healthcare providers each other. Later, the family meeting proceeded in the conventional manner. Following each of the family meetings, the family members were requested to fill the surveys on the family members’ levels of satisfaction.
Sample size estimation
Assuming a variance of approximately 64 for the overall score for satisfaction in each group, we included at least 40 participants in each group to detect a difference of at least five points in the median overall satisfaction score between the pre- and post-intervention groups with a power of 0.80 and an α of 0.05. The predicted difference between the groups was estimated based on the pilot-testing results.
Data were reported as numbers (percentages) or medians (interquartile ranges) as appropriate. To examine whether there were any differences between the levels of satisfaction of the family based on the questionnaire items before and after the intervention, the Mann–Whitney U test was used. Comparisons between the family’s baseline characteristics were performed using the χ2 test. Cronbach’s α was adopted to examine internal consistency among the items. All tests were two-tailed and a P value of < 0.05 was regarded as statistically significant. All statistical analyses were performed using the Statistical Package for Social Sciences (SPSS, version 22.0, SPSS Inc., Chicago, IL).