Study design
The study was a multicentre cross-sectional survey. Participants’ demographic characteristics, well-being, quality of life, quality of work-life, and quality of care were obtained using four standardized questionnaires.
Study locations
Southwestern Nigeria is made up of six states. We purposively selected four publicly funded tertiary hospitals in Ogun, Osun, and Oyo states based on the socio-cultural, environmental, political, and socioeconomic similarities of those states. It was assumed that the three states were representative of the others. The selection criteria for the hospitals were: (i) being publicly funded, (ii) being a referral centre with a tertiary level of care, and (iii) having a bed capacity greater than 500. The included hospitals were FMCA and OOUTH Sagamu, in Ogun State; OAUTH, Ile-Ife, Osun State; and UCH Ibadan, Oyo State.
Study participants
The designated health professionals were medical practitioners (physicians, surgeons, dentists), nurses, physiotherapists, pharmacists, medical laboratory scientists. Participants were proportionally recruited based on the population size of each hospital using the formula: ([Z2P{1 – P}]/e2) / (1 + ([Z2P{1 – P}]/e2N)), where proportion (P) = 0.5, margin of error (e) = 0.05, Z-score = 1.96, populations (N) = 566 (OOUTH), = 2000 (FMCA), = 3000 (UCH), and 1490 (OAUTH). Therefore, the minimum sample size for OOUTH, FMCA, UCH and OAUTH were 229, 323, 341, and 306, respectively, giving a total of 1199. In anticipation of incomplete survey response, 1,600 participants were recruited.
Study instruments
A biodata form and four standardized questionnaires were used for data collection. The biodata form was used to obtain information on participants’ demographic variables such as gender, age, years of practice, highest educational qualification, designation, appointment type, work schedule, average weekly work hours, and practice location. Participants’ well-being was assessed using the WHO-5 well-being index and the personal well-being index (PWI) scale [25-26]. The WHO-5 is a five-item questionnaire that assesses participants’ feelings about components of healthy living in the past two weeks on a 6-points Likert scale (score 0 to 5). Each participant’s responses were summed (range = 0 to 25) and converted into percentage scores. The WHO-5 has been reported to be valid and reliable in a systematic review of 213 studies that applied the instrument in diverse settings, with average sensitivity and specificity of 0.86 and 0.75, respectively [26]. The PWI is an eight-item valid and reliable questionnaire that assesses participants’ satisfaction with their life, health, life achievements, relationships, safety, community, future security, and spirituality on an 11-points (score 0 to 10) Likert scale [25]. Each participant’s responses were summed and converted into percentage scores [25]. The PWI has been reported to have good psychometric properties: validity, reliability, and sensitivity [27]. The internal consistency measured with Cronbach’s alpha ranges between 0.86 and 0.89 [28]. The work-related quality of life (WRQoL) questionnaire was used to assess participants’ quality of work-life. The 24-item questionnaire contains six domains: general well-being, home-work interface, job-career satisfaction, control at work, working conditions, and stress at work [29]. The WRQoL contains 21 positively worded questions and three negatively worded questions (items 7, 9, and 19). The questionnaire asked the extent to which a participant agreed to each of the items as it related to their work-life on a 5-point Likert scale, 1 = strongly disagree to 5 = strongly agree. The WRQoL questionnaire was found valid, reliable, and consistent among a cohort of health professionals, overall scale reliability was 0.91 with good subscale reliabilities ranging from 0.76 to 0.91 [17]. Furthermore, the participants’ quality of care was obtained using the Quality of Care (QoC) questionnaire [30]. The 22-item questionnaire was subdivided into two domains: positively worded person-centred care (12 items) and negatively worded discordant care (10 items). The questionnaire inquired on how frequently the listed items had occurred while the participant related with patients in the last six months. Responses were on a 6-point Likert scale, 0 = never to 5 = always. The QoC questionnaire has been shown to be valid and reliable, with Cronbach’s alpha of 0.86 and 0.74 for the person-centred care and discordant care subscales respectively.
Procedure for data collection and storage
Ethical approval was independently obtained from the health research ethics committees of: The University of Lethbridge, Alberta, Canada (protocol number 2021-053); Federal Medical Centre, Abeokuta, Ogun State, Nigeria (FMCA/470/HREC/01/2021/07); Olabisi Onabanjo University Teaching Hospital, Sagamu, Ogun State, Nigeria (OOUTH/HREC/415/2021AP); University of Ibadan/University College Hospital, Ibadan, Oyo State, Nigeria (UCH-UI/EC/21/0125); and Obafemi Awolowo University Teaching Hospital Complex, Ile-Ife, Osun State, Nigeria (ERC/2021/11/03). The research was conducted in accordance with the ethical principles guiding human subjects’ research and informed consent was obtained from all participants. The questionnaires were distributed to health professionals within the selected hospitals with the help of research assistants. Health professionals were included in the study if they had worked in one of the selected facilities for at least six months and were willing to read and sign a written informed consent form before responding to the survey. The questionnaire was self-administered and returned to a research assistant after completion. The primary investigator’s number was boldly printed on the survey to facilitate the return of questionnaires in the case of participants who failed to submit theirs immediately. Data were extracted from the questionnaires and transferred to already coded SPSS spreadsheet in designated computers at the four study locations. Individual datasets were merged into a final anonymized dataset, password encrypted and saved to the cloud.
Data analyses
Data were analysed using SPSS 27.0 version software (SPSS Inc., Chicago, Illinois, USA). The dataset was cleaned of missing variables, all entries with more than 25% missing variables in a domain were deleted for the domain. The data were summarized using descriptive statistics: frequency (percentage) and mean (standard deviation). Participants’ well-being, QoWL, and QoC scores were summated in separate columns and converted to percentage points, this was in line with the rubric provided by the instrument developers and the general approach for analysis of Likert scale data [31]. We obtained continuous normally distributed variables for each of the outcomes (skewness < 3.29). We dichotomized the QoC scores into poor and good QoC using a 75% cut-off for good QoC. Inferential statistics included Chi-square test for differences in the QoC levels across the demographic characteristics; Pearson’s correlation among the outcomes; independent samples t-test analysis for differences in the mean well-being index and QoWL among people classified to have provided poor and good QoC. The domain validity and reliability of the aggregate instrument were obtained using Cronbach alpha and Intraclass correlational statistics. We completed a Confirmatory Factor Analysis (CFA) using Maximum Likelihood with Varimax-orthogonal rotation to determine the new domain membership of the items in the aggregate questionnaire. A structural equation modelling through path analysis for the relationship between QoC with other study outcomes was completed using SPSS Analysis of Moment Structures (AMOS). A maximum likelihood estimation procedure was used to estimate the coefficients. The model fitting was assessed using a Chi-square goodness of fit test as well as the comparative fit index (CFI), the root-mean-square error of approximation (RMSEA), and Tucker-Lewis index (TLI). The measurement of association was interpreted under the standardized mode, using the odds ratio and confidence intervals of 95%.