Study design and reporting
A descriptive We used a cross sectional study design was used to determine the practice, perceived barriers and motivating factors to medical-incident reporting at MRRH in Southwestern Uganda. This study design was selected to give a snap shot of MIR among HCPs at a tertiary health facility. In this study, individual reporting of medical incidences was assessed by means of the questionnaire.
Study setting and, population, sampling and sample sizeparticipants’ selection
This study was conducted at Mbarara Regional Referral Hospital (MRRH), one of the 13 Regional Referral Hospitals in Uganda. MRRH is a government owned referral hospital that started in the 1930s as a nursing school for the South-western Ankole region of Uganda. It then arose to be a district hospital in 1950, and consequently to a regional referral hospital a status that was granted in 1990s. Presently, the hospital serves a population of over four million people mainly from the districts of Mbarara, Bushenyi, Ntungamo, Kiruhura, Ibanda,Buhweju, Rubirizi, Mitooma and Isingiro. The hospital also receives patients from Kabale, Masaka, Fort Portal and neighboring countries like Rwanda, Tanzania, Burundi and Democratic Republic of Congo. MRRH comprise wards, these are: Accident and Emergency, the Intensive Care Unit (ICU), Surgical, Medical, Obstetrics and Gynecology, Pediatrics, Psychiatry and Tuberculosis (TB) wards. Also, MRRH has a private wing for both male and female patients, laboratories, operating theaters, the Eye center, Cardiology clinic, Oncology unit, Physiotherapy unit, Immune Suppressed Syndrome (ISS)/HIV clinic, Maternal and Child Health unit, and General out-patients. MRRH was purposively selected because it is the biggest hospital in the southwestern region offering comprehensive subspecialties (clinical and diagnostic services). and cares to an estimated 4 million people. In this study health workers that include: doctors, clinical officers, nurses, dispensers, laboratory personnel, pharmacists and anesthetists respectively were interviewed. Only 158 of 197 (80.2% response rate) health care workers verbally consented to take part in the study, the respective number of questionnaires administered to each category are:. These comprised of nurses (66.5%N = 105), laboratory personnel (N = 7), midwives (N = 12), doctors (N = 14), clinical officers (N = 12), anesthetist (N = 31.9%), pharmacist (N = 21.%), theatre assistant (N = 21.3%) and dispenser (N = 10.6%) respectively. The study population comprised HCPs at MRRH who had served for at least a year. Those who at the time of the study were away for leave were excluded. Census sampling was used in which all the HCPs were considered.
Data instrument and collection
We gathered data using a self-administered questionnaire (Appendix I), and an additional note was attached with each questionnaire which defined the aim of the research. The questionnaire comprised closed ended questions measured on a four-level Likert scale from 1–4 (Strongly agree, Agree, Disagree, and Strongly disagree, respectively). The questionnaire was developed from existing literature on earlier studies [8, 10–15]. It was structured into: Medication, surgical, and preventive incidences/errors ever witnessed at MRRH; individual practice of incidence/error reporting; motivating factors; and perceived barriers towards MIR. Two members of the research team assisted by two senior HCPs who had more than 10 years of experience in patient safety research assessed the questionnaire for face and content validity. Adjustments were made based on the need until such a point when it was deemed satisfactory, that it was even further pretested on thirty HCPs prior to administration to the study participants. The research team liaised with the staff on duty based on the existing monthly rostaroster, and a convenient time to consent and administer a questionnaire was chosen based on the HCPs availability and work schedule. The research team utilized the existing duty rooms for seeking participant’s consent during which the purpose of the study was explained to participants, and thereafter, a questionnaire was given out to be completedread and fill in as a member of the research team was nearby in case one needed clarification for such a question(s) found difficult to understand. These were rephrased for a clear understanding by the interviewee.
Study variables, population, recruitment criteria and sampling techniques
We extracted data on the types of incidents, motivating factors, and perceived barriers to MIR respectively. The dependent variable was medical incident reporting (MIR), coded and was measured as a binary outcome (Yes-reported or No-not reported). A health worker who was directly or indirectly affect by an incident or promptly reported so that appropriate action was taken was considered to have addressed a medical incident. Conversely, any health worker that observed or was injured or witnessed an injury to a patient or staff but did not report the incident were considered non-compliant. The study population comprised HCPs at MRRH who had served for at least a year. Those who at the time of the study were away for leave were excluded. Census sampling was used in which all the HCPs were considered.
Quality control
The questionnaire was pre-tested among thirty consented HCPs within two health centres (HCs) in Mbarara Municipality, namely: Mbarara Municipal Council HC IV, and Kakoba HC III. Quality of data was guaranteed through self-administered questionnaires to avoid any bias, and the research assistants were available and within close range to the respondents, in case the latter needed clarity. The research assistants checked and ensured the completeness of the questionnaire.
Data analysis
The data gathered was entered into SPSS (statistical software for social scientist), cleaned, then analyzed at univariate level of analysisfor univariate, and bivariate at 5% significance level. At univariate analysis, and presented as descriptive statistics in the form of frequencies and percentages were used for categorical variables. To establish which factors contributed to MIR, the principle component analysis (PCA) was used and any variable with a component loading of greater than 0.60 was considered to be correlated withsignificant influence MIR at MRRH. While the Bartlett’s test (approximation of the Chi-square test) was used to test whether the variances in the sample were equal across the groups. Only associations with probability value (P-value) less than 5% were considered to be statistically significant in determining MIR at MRRH.
Ethical considerations
As this study did not involve human sampling or clinical procedures/intervention otherwise considered by Uganda National Council for Science and Technology to interfere with humans [22], it received an exemption of ethical clearance from Clarke International University (Formerly, International Health Sciences University). However, administration approval was obtained and this was presented to the Director MRRH, who then authorized the conduct of this research. In this study, HCPs consent was obtained from all participants. Their responses were treated with utmost confidentiality and not linked to participant’s identity.