This present study is a cross-sectional, descriptive-analytical study that was conducted to investigate the reasons for delivering futile life-saving interventions from the perspective of physicians and nurses working in hospitals affiliated to TUMS in 2018-2019 and to compare their perspectives. The study population was all physicians and nurses working in hospitals affiliated to TUMS, resulting in participation by 128 physicians consisting of new medical graduates, residents, specialists, fellows and subspecialists in the fields of anaesthesia, haematology-oncology, cancer surgery, internal medicine, general surgery and paediatrics, as well as 121 nurses from Intensive Care (ICU), haemato-oncology and cancer surgery departments.
After obtaining research and ethics approval (Code: IR.TUMS.MEDICINE.REC.1397.706), the researcher, in collaboration with ward managers, met with the study participants and after stating the study’s objectives, recruited them into the study. Then, a questionnaire was distributed among eligible participants. Completion of the questionnaire was considered equivalent to consent to the study while those who did not complete the questionnaire were deemed to have declined.
Sampling was via a multi-stage method. First, a list of hospitals affiliated to TUMS was collated. Then, hospitals housing more wards related to end-of-life care such as ICU, haemato-oncology and cancer surgery were selected for the study. Then in these hospitals, wards belonging to ICU, haematology-oncology, cancer surgery, internal medicine, general surgery and paediatrics were selected for sampling. In the next step, convenience sampling was performed among physicians and nurses from each ward according to the inclusion criteria. The inclusion criterion was employment of at least 2 years.
To calculate the sample size, 20% of the data of each group of physicians and nurses were randomly selected and its indicators (as a pilot study) were used to estimate the sample size of the main study. According to the pilot study indicators and the use of the formula below the sample size required for each group (of physicians and nurses) was 65 participants (alpha 5%, beta 20%, mean and standard deviation of physicians 2.45 ± 1.18 and nurses 1.96 + 0.82). To increase the study accuracy more participants (128 physicians and 121 nurses) were recruited.
To collect data a questionnaire was used to assess the causes of delivering futile life-saving treatments and interventions, designed and validated by Aghabarary et al (2017). The original questionnaire consisted of 35 items, of which 25 were chosen as these were most relevant to participants and the objectives of the study. In the study, the reliability of the questionnaire was evaluated and confirmed using Cronbach's alpha method. The coefficient for all three domains [detailed below] and the whole instrument was calculated at 0.75.
The questionnaire consisted of two parts. The first part collated demographic information and the second was related to the reasons for providing futile life-saving treatment and interventions in three domains of patient/family demand (Q1-9), personal reasons of healthcare staff (Q10-12, Q14-16, Q24-25) and, organizational/infrastructural constraints (Q13, Q17-23). Each item was scored by a 5-point Likert scale with options for “strongly agree” (1), “agree” (2), “no opinion” (3), “disagree” (4) and “strongly disagree” (5). The higher the points, the greater the level of disagreement with the statement. The percentage of agreement of participants with each item was calculated by adding the percentage of “completely agree” and “agree” options.
Data were collected in the nurse’s group in person. In the physician’s group, both in-person and in-absentia (via an electronic version of the questionnaire distributed via the Telegram® platform) methods were utilised to collect data. Data were analysed using the SPSS-16 software. Results were presented in terms of absolute and relative frequencies, mean and standard deviations. Also, data related to participants' agreement with each cause of futile life-saving interventions were expressed as percentages. Comparison between the physician and nurse group was performed using the Chi-Square test. A p-value less than 0.05 was considered to reflect a significant difference.
A regression analysis was used to compare the responses of physicians to nurses within each of the three domains of the questionnaire.as well as test which domain was considered the most significant cause of futile intervention within each group of physicians and nurses. For this, after the scores of each domain were standardised with attention to a normal distribution, a linear regression analysis was performed where the dependent variable was the total score obtained from the questionnaire in that group and the independent variable was the (standardised) score obtained from the domains. Thus, two regression models for physicians and nurses were separately calculated, with R2 indicating the percentage change in the total score of each group due to each domain in that group. The standard deviation of R2 was calculated with a 95% confidence interval using the formula SD R2 = √ ((1-R2) / (n-2). If the confidence intervals of the other two domains did not overlap the point estimation of the third domain, it was considered a significant difference in the effect of that domain on changes of the total score. Using the same method the effect of [the other two] domains on the total score was analysed.
In this study, all necessary ethical considerations were observed. Physicians and nurses’ participation in the project was voluntary, and necessary information was provided before completing the questionnaire. Returning questionnaires was understood to indicate consent to participate in the study. Questionnaires were completed anonymously, and participants' data was kept strictly confidential.