Background: Adverse medical events (AMEs) are threats to delivery of quality healthcare services, particularly in resource-poor settings such as Ghana. In sub-Saharan Africa, 30% of deaths are attributed to AMEs and a significant proportion of these events are not reported. This study explored personal experiences of nurses with AMEs and the constraints to reporting them. Methods: This is a descriptive cross-sectional study among professional (n = 133) and auxiliary (n = 88) nurses in a regional referral hospital in northern Ghana. A test for differences in experiences of professional and auxiliary nurses was done using Wilcoxon Mann-Whitney test. Ordered logistic regression analysis (proportional odds ratio models) and probit regression were used to ascertain the determinants of staff’s knowledge on AMEs and the odds of exposure, respectively. Results: Overall, knowledge and awareness level on AMEs was average (mean = 3.1 out of the five-point Likert scale of 1 = “Very poor” to 5 = “Excellent”). Knowledge levels among professional nurses (mean = 3.2) were relatively higher than those among auxiliary nurses (mean = 3.0), (p = 0.006). The predominant type of AME experienced was wrongful documentation (n = 144), and the least experienced type was wrong transfusion of blood and/or intravenous fluids (IVF) (n = 40). Male staff had higher odds of experiencing medical errors relative to female staff, OR = 2.39 (95% confidence interval (CI), 1.34-4.26). Inadequate logistics was the most perceived cause of AMEs. Knowledge on types of AMEs was significantly associated with gender of the respondents, OR = 1.76 (95% CI, 1.05-2.94); moreover, male staff had higher odds of knowing AME post-exposure action than female staff, OR = 1.75 (95% CI, 1.04-2.93). Conclusion: Knowledge levels of nursing staff on AMEs were generally low, and even though exposures were high they were not reported. There is the need to integrate AME modules into the pre-service and in-service training curricula for nurses to enhance their knowledge on AMEs; reporting registers for AMEs should be made available in clinical sites and staff incentives given to those who report AMEs. Lastly, protocols on AMEs should form part of the quality assurance value chain for health facilities to promote compliance.
This is a descriptive cross-sectional study conducted among different cadres of nursing staff in a major regional referral hospital in the Upper West Region of Ghana. The Upper West Region (UWR) is one of the administrative regions located in northern Ghana. The region shares boundaries with Burkina Faso to the north, Upper East region to the east, Ivory Coast to the west, and Northern region to the south. UWR covers a geographical area of approximately 18,478 km2 and constitutes about 12.7% of the total land area of Ghana; the region is located in the guinea savannah vegetation belt and has a population of 702,110 (341,182 males and 360,928 females) [11]. The study hospital is a major referral hospital for lower level hospitals within UWR, parts of Northern region and neighboring Burkina Faso. The hospital has a total bed capacity of 181 at the time of conducting this study. Services rendered in the facility include general medical services, ante-natal, post-natal, and maternal care. Other service components include specialist care such as ear nose and throat (ENT), dental, laboratory, and physiotherapy. The study population included all cadres of nursing staff on permanent employment. The regional hospital has a total staff population of 682 at the time of conducting this study. Out of this number, 221 were paramedics, 114 casual workers, and 347 nurses of various categories (Upper West Regional: Hospital Administrative Records, 2018; unpublished). Thus, nursing personnel represented approximately 51% of the total workforce in the hospital. The sampling technique was a census of all professional and auxiliary nurses across all the units of the hospital. Since the nursing staff population was 347, the entire population served as the target sample size (n = 347) for the study. However, three (3) extra questionnaires were printed to take care of instances where staff misplaced their questionnaire. This strategy was precautionary because not all questionnaires were retrieved on the day of visit due to busy schedules of some respondents. The study included nurses of all categories such as professional and auxiliary nurses (i.e., staff licensed by the Nursing and Midwifery Council of Ghana). Only staff on permanent appointment were eligible for inclusion in the study. Also, staff who worked for at least six (6) months on the day of visit were included in this study to obtain data that is reflective of the true experiences of respondents. Staff on post-retirement contract, student nurses, or nurses on rotation/internship were equally excluded. A structured questionnaire, comprising of both closed and open-ended questions, was used for the data collection. Since all the target respondents were literates, the questionnaires were largely self-administered and later followed up by the researchers for retrieval. The data collection instrument comprised of four (4) main sections namely: section A (socio-demographic characteristics and work history), section B (experiences and exposure to AMEs), section C (perspectives on causes of AMEs), and section D (reporting of AMEs). Some of the questions on experiences/exposure to AMEs were dichotomized into “Yes” and “No” while others were ranked on a four-point Likert scale as follows: 1 = “strongly disagree,” 2 = “disagree,” 3 = “agree,” and 4 = “strongly agree.” Questions on knowledge levels on AMEs were measured on a five-point Likert scale of 1 = “very poor,” 2 = “poor,” 3 = “average,” 4 = “good,” and 5 = “excellent.” All 22 Likert scale items were tested for scale reliability, and mean Cronbach’s alpha was found to be 0.81, which is acceptable. The questionnaire was subjected to peer reviews and one pre-testing to promote its validity and reliability. The pre-testing did not lead to changes in the questionnaire, except correction of few typographical mistakes. Moreover, design of the questionnaire was guided by the research objectives and reviewed literature. Data was analyzed using the STATA statistical analysis software (version 12.0, StataCorp, College Station, TX, USA). Field data was first captured with Microsoft Excel, cleaned and coded before exporting to STATA for analysis. Chi-square (χ2) and Fisher’s exact tests were used for the bivariate analysis of categorical data as appropriate while summary statistics on continuous variables were analyzed using independent Student’s t test. A test for differences between professional and auxiliary nurses on the Likert scale items was determined using Wilcoxon Mann-Whitney test. Un-rotated factor analysis was conducted on the various Likert scale items to aggregate the various Likert scale items into similar components. Thus, seven items on staff knowledge on AMEs were factor-analyzed and three were retained, namely “types AMEs,” “action(s) after AME experience,” and “ability to recognize an incidence of AME.” Questions on staff perspectives on the causes of AMEs were also factor-analyzed to arrive at five retained factors out of ten factors. The five retained factors were “poor communication,” “inadequate staff,” “poor management of previous AMEs,” “inadequate skills on AMEs,” and “inadequate motivation.” Finally, staff perspectives on barriers to reporting AMEs and the corresponding constraints were factor-analyzed and two factors retained out of five. The retained factors were “access to incidence report book” and “lack of clear reporting system.” Following the factor analysis, ordered logistic regression analysis (proportional odds ratio models) was performed to ascertain determinants of the key outcome variables of interest. The main outcome variables of interest for the ordered logistic regression were the factor-analyzed proxies on knowledge, perceived causes, and constraints to reporting AMEs. Independent variables in the logistic regression were staff age (18–30 years = 1, otherwise = 0), education (first degree = 1, otherwise = 0), gender (male = 1, otherwise = 0), marital status (married = 1, otherwise = 0), religion (Christianity = 1, otherwise = 0), and work experience (5 years or less = 1, otherwise = 0). All independent variables were dichotomized to create uniform reference points (dummies) and enhance ease of interpretation of the findings. Determinants of personal experience/exposure to AMEs were explored using probit regression. The main outcome variables were five factor-analyzed components on personal experiences with the various AMEs (yes = 1, no = 0). The five factor-analyzed components were needlestick pricks, equipment-related injuries, falls from slippery floor, medication errors, and falls from height. The independent variables in the probit regression are the same as those used for the logistic regression described earlier. All independent variables were tested for multicollinearity prior to their inclusion in the regression models, and the mean variance inflation factor (VIF) was 1.26. None of the independent variables had a VIF above 10.0 necessary for exclusion from the regression models. Statistical significance was set at 95% for all analysis.
N/A