Background: In countries with high maternal and newborn morbidity and mortality, reliable access to quality healthcare in rural areas is essential to save lives. Health workers who are satisfied with their jobs are more likely to remain in rural posts. Understanding what factors influence health workers’ satisfaction can help determine where resources should be focused. Although there is a growing body of research assessing health worker satisfaction in hospitals, less is known about health worker satisfaction in rural, primary health clinics. This study explores the workplace satisfaction of health workers in primary health clinics in rural Tanzania. Methods: Overall, 70 health workers in rural Tanzania participated in a self-administered job satisfaction survey. We calculated mean ratings for 17 aspects of the work environment. We used principal components analysis (PCA) to identify groupings of these variables. We then examined the bivariate associations between health workers demographics and clinic characteristics and each of the satisfaction scales. Results: Results showed that 73.9% of health workers strongly agreed that they were satisfied with their job; however, only 11.6% strongly agreed that they were satisfied with their level of pay and 2.9% with the availability of equipment and supplies. Two categories of factors emerged from the PCA: the tools and infrastructure to provide care, and supportive interpersonal environment. Nurses and medical attendants (compared to clinical officers) and older health workers had higher satisfaction scale ratings. Conclusions: Two dimensions of health workers’ work environment, namely infrastructure and supportive interpersonal work environment, explained much of the variation in satisfaction among rural Tanzanian health workers in primary health clinics. Health workers were generally more satisfied with supportive interpersonal relationships than with the infrastructure. Human resource policies should consider how to improve these two aspects of work as a means for improving health worker morale and potentially rural attrition. Trial registration: (ISRCTN 17107760).
This analysis was conducted within a large cluster-randomized maternal and newborn health quality improvement study (ISRCTN 17107760). 24 government-managed primary healthcare facilities (dispensaries), in four districts of rural Pwani Region, Tanzania (Bagamoyo, Kibaha Rural, Kisarawe, and Mkuranga) were selected for inclusion in the study (Figure 1). To be included, facilities had to be government-managed, have at least one medically trained staff member (e.g., clinical officer or nurse), and be actively providing delivery services. From the 62 eligible facilities, the 6 primary care clinics from each district with the highest volumes of deliveries in the 6-month period from January to June 2011 were selected for inclusion. Map of study facilities, Pwani Region, Tanzania. All health workers working in the study facilities were invited to participate in the satisfaction survey. If health workers were not available on the day of data collection, the study team arranged for an additional visit to the facility in order to increase participation. The self-administered survey was adapted from the Revised Nursing Work Index [25]. The survey includes a 17-question index addressing aspects of the work environment, asking health workers to state their agreement on a 4-point Likert scale. Additionally, we conducted an assessment of the 24 primary care clinics where the health workers were employed. This was done using a structured questionnaire adapted from the needs assessment created by the Averting Maternal Death and Disability Program and the UN system that has been previously used in more than 30 countries, including Tanzania [26]. The survey included questions regarding human resources, infrastructure, and services available as well as a record review of the volumes of services provided and was answered by the most senior provider available on the day of data collection. Data were collected from December 5, 2011, to May 15, 2012, by three teams of three data collectors. Data collectors were trained for one week in ethical data collection and conducting interviews. The project was approved by ethical review committees at Columbia University (United States), Ifakara Health Institute (Tanzania), and National Institute for Medical Research (Tanzania). Written consent was obtained from each participant. The survey administered to health workers included 17 questions related to aspects of the work environment. An additional three questions related to overall job satisfaction were asked, namely, how satisfied they were with their own job, how satisfied other people are in similar jobs, and, finally, how much they would like to continue to work for this facility. Based on the literature we identified a range of determinants of health worker satisfaction: health worker demographics, health worker length of employment, facility size, facility infrastructure, and services provided by the facility. We assessed health worker’s age, sex, cadre, and whether they were full or part-time. We categorized the time the health worker had worked in the facility at more than two years based on an inflection point seen in the data. To assess the workload of the facility we looked at the average monthly facility deliveries in 2011 and the average monthly outpatient visits for 2011. Where data were missing for a month in a facility we substituted the average of the contiguous months (data were missing for 2.3% of facility-months). If data were missing from two consecutive months (two instances), we created an average using the months where facility data were available. We also looked at the number of outreach visits the facility conducted in the 90 days before interview and the number of health workers at the facility. We assessed the facilities’ performance of basic emergency obstetric care (i.e., number of obstetric signal functions in the past three months, such as administering uterotonics or conducting newborn resuscitation), as a measure of the complexity of obstetric services provided in the facility. In order to assess the effect of the availability of equipment, supplies, and drugs on health worker satisfaction, we created an index. We utilized Tanzanian government guidelines, findings from the literature, and expert opinion to develop a list of essential obstetric equipment, supplies, and drugs [27,28]. We then calculated a summative score, where a single point was given for each item available and functioning on the day of assessment. We further assessed the infrastructure of the facility through the availability of electricity, clean water, and whether or not the facility had received an upgrade or renovation in the past year. We assessed the extent of supervision through the number of managerial meetings and supervisor visits in the past 90 days. Finally, because management of health facilities occurs through district-based teams, we assessed differences across the four districts in our study. Data were entered and variables were examined for missingness and outliers. Descriptive statistics were calculated for health worker and facility-level characteristics. We created a binary variable that grouped health workers into “strongly agree” versus all others, for health workers’ assessment of general job satisfaction. We conducted a principal components analysis (PCA) of the 17 questions related to aspects of the work environment. PCA identifies underlying components that are described by the index through identifying questions that strongly correlate. Cattell’s Scree test graphically demonstrated that a majority of the variation in satisfaction was due to two components. We therefore extracted two components and used varimax rotation to simplify interpretation of the factors by allowing each variable to load strongly on only one factor. All 17 questions from the original index were maintained in the PCA. We created scores for each of the subscales using the sum of the subscale regression weights (factor loadings) multiplied by the healthcare worker’s response for each question [29]. We standardized the resulting subscales to aid in interpretation. We calculated Cronbach’s α to assess internal consistency of the subscales. We then examined the association between the satisfaction subscale scores and each of the characteristics of the health workers and health facilities using bivariate linear regression analyses. We further explored correlation between dependent variables. Data analysis was conducted using Stata version 12.1 (StataCorp, 2012, College Station, TX, USA).
N/A