Low levels of knowledge and practice of occupational hazards among flower farm workers in southwest Shewa zone, Ethiopia: a cross-sectional analysis

Flower farms based cross-sectional survey.

This study was conducted in Southwest Shewa zone, National Regional State of Oromia, Ethiopia, from March to April 2017. Southwest Shewa zone is one of the zones of the National Regional State of Oromia, Central Ethiopia. Wolliso, the capital of the zone, is an ideal place for investment activities, particularly agroindustry. The town is 114 km from Addis Ababa, the capital of Ethiopia, and it has 12 districts and town administration. At the time of data collection, there were over 25 industries in the area, of which five were flower farms. Moreover, during data collection period, there were about 1500 employees working in the flower farm industries.

All workers in the flower farms in Southwest Shewa zone were the source population, whereas those who met the inclusion criteria and available during data collection were the study population.

Employees who had worked 3 months and above prior to the data collection were included, while those who were sick, on annual and maternity leaves were excluded.

Epinfo program version 7 was used to calculate the required sample. We derived the levels of knowledge and safety practice of occupational hazards from previous studies [6, 21, 22]. Accordingly, proportions of 72.9 and 76.3% with a Confidence Interval (CI) of 95 and 5% margins of error were presumed and 268 and 248 samples were calculated for the levels of knowledge and safety practice, respectively. Similarly, for each specific objective, three associated factors including training on workplace conditions (43.1% prevalence and 3.5 Odds Ratio (OR) with a CI of 95%), not using PPE (30.2% prevalence and 4.065 OR with a CI of 95%), and not using shower (43% prevalence and OR 0.569 with a CI of 95%) were taken from the aforementioned studies to attain 118, 136 and 428 samples, respectively. Finally, we took the largest sample size (428) to ensure the adequacy of the sample for statistical power. We assumed a 10% for none response rates which gave the final sample of 471.

The stratified sampling technique was employed, considering that the populations in each selected industry were heterogeneous. The five floriculture industries were included purposively to attain the required sample. A proportional allocation was used to derive sample from each stratum. Finally, using the lists of workers’ identification numbers provided them by the industries, we applied a computer-generated random number to reach each participant.

Data were collected by use of a structured and interviewer-administered questionnaire. We developed the questionnaire after a meticulous review of published works [17, 23‐27]. The questionnaire has four sections. The first section contains socio-demographic characteristics including sex, age, educational status, profession, marital status, and work experience. The second, included questions related to the level of knowledge on occupational hazards. The third and fourth sections provided detailed information on safety practices and factors affecting the levels of knowledge and safety practice (dependent variables), respectively. To assess the level of knowledge, detailed lists of knowledge questions (14 items such as knowledge on safety communication (safety labels, symbols, pictograms, guidelines), material safety data sheets (MSDS), healthy effects of hazards, and routes of exposures) were presented. Responses to the questions are coded such that correct answers (Yes) scored one and incorrect (No) zero. Knowledge score was then categorized as not knowledgeable for < 50% and knowledgeable for > 50% correct responses from the overall scores [27].

A questionnaire containing 16 items was administered to assess the level of safety practice and the scores were categorized as poor for < 50% and good practice for > 50% correct (Yes) responses out of the total scoring [26]. The questionnaire was first prepared in English and translated to Afan Oromo (the local language) and retranslated to English by independent language experts to verify its consistency. Finally, the Afan Oromo version of the questionnaire was interviewer-administered to the participants at their work sites. Moreover, we designed a standard checklist to evaluate workplace hazards and observe employees’ onsite safety practice.

To ensure the quality of data, first, a pretest was conducted on 5% of the sample (48 workers) in flower farm at Sebeta flower farm, which have similar characteristics as those included in the final survey. Based on the test results, we reduced the number of questions (without changing what was intended to be measured) to minimize the time needed for interviews, and we modified ambiguous questions. Secondly, training and orientation was given to data collectors (2 females, 4 males) and two supervisors on issues relating to the objectives of the study, confidentiality of data, consent and appropriate time for data collection.

The collected data were entered into Epi-info version 7 tocleanand code the collected data. Double entry was performed with 10% data to verify errors that couldoccur during the entireentry process. The data were analyzed with SPSS program version 20. The findings were presented using descriptive statistics includingfrequency tables, graphs, percentages, means with standard deviations. Knowledge and practice questions were marked with ‘0’ for ‘No ‘and ‘1’ ‘Yes’ response.

We checked the reliability of the knowledge and practice items using a Cronbach’s Alpha Coefficient. As such, the knowledge questions yielded a Coefficient of 0.78, while that of the practice 0.81. It was previously shown that a Cronbach’s Alpha Coefficient of a given instrument is considered reliable if it is > 0.65. Before running the multivariable linear regression analysis in the final model, linearity, normality, outliers, autocorrelation, multicollinearity and independence of errors/residues of the variables were also examined. The multicollinearity test was done using variable inflation factor (VIF) and all variables showed VIF < 5. A bivariate linear regression analysis was performed to examine associations of each independent variable and knowledge and practice separately. Independent variables with < 0.2 p-values in this type of analysis were exported to the multivariable linear regression analysis to control effects of potential confounders. We set the significance of associations at < 0.05 p-values, while adjusted odds ratio (AOR) with 95% confidence interval (CI) was used to determine the strength of associations.

A total of 451 flower farm workers were interviewed with a response rate of 95.7%. Seventy-two percent (N = 325) of those surveyed participants were females. Mean age of the participants was 24.1 (SD + 6.5) years. Of the interviewees, 86.3% (N = 389) of them were Oromo, 8.0% (N = 36) Gurage, and 5.7% (N = 26) Amhara. The majority, 54.8% (N = 247) of the participants were urban, while 45.2% (N = 204) were rural residents. More than half, 46.8% (N = 211) of the interviewees were temporary and the remaining were permanent workers. With respect to monthly salary, 58.9% (N = 266) of the workers earned less than 1500 ETB (50 USD) per month. More than half, 58.1% (N = 262) worked in closed and hot conditions, whereas 22.4% (N = 101) of the participants worked in open and hot environments (Table 1).

Figure 1 shows that almost all, 92.2% (N = 416) of the respondents had no regular supervisions and communication on safe work procedures; 86.5% (N = 390) worked without health and safety instructions, symbols and pictograms; 82.7% (N = 374) reported there were no programs for health and safety at work.

Of the participants, 39.2% (N = 177) [95% CI (34.8, 43.9)] showed that they had a good level of knowledge (knowledgeable), whereas 60.8% (N = 274) [95% CI (56.1, 65.2)] of the participants scored poor (not knowledgeable) on occupational hazards. The mean knowledge score was 7.15 (SD + 1.896). Of the participants, 50.6% (N = 228) revealed they had no information about the hazards associated with their current jobs. A lower proportion, 38.6% (N = 174) of the sampled workers indicated that their sources of information about occupational hazards was experienced co-workers; 25.1% (N = 97) guidelines provided by the manufacturer, and 24.8% (N = 112) trainings. Out of the respondents, about 42.4% (N = 191) pointed out they knew the presence of chemical hazards, while 36.8% (N = 166) and 26.6% (N = 120) of them knew the presence of physical and psychosocial hazards, respectively. Table 2 depicts knowledge responses of the participants towards occupational hazards.

The overall level of safety practice was 26.6% (N = 120). About 68.3% (N = 308) of the respondents explained they correctly used health and safety communication; 57.2% (N = 258) reported sudden occurrences/events; 53.2% (N = 240) properly used information and work materials provided by the industries and 53.0% (N = 239) of the surveyed workers stated that they used health and safety instructions. Regarding the use of personal protective equipment (PPE), 60.1% (N = 271) of the respondents indicated they did not use it regularly. Few workers, 4.4% (N = 20) observed they used full body protection, while 36.1% (N = 163) reported they used gloves (Fig. 2). Moreover, 61.4% (N = 277) of the participants reported that waste emitted from the flower farms was not properly disposed; 60.8% (N = 274) of the workers did things not recommended such as eating, drinking, not washing hands after work, smoking, taking and washing PPE at home, used empty chemical, fertilizer and other raw material containers for water storage at home and selling. Table 3 describes practice responses of the participants towards occupational hazards.

We observed majority of the workers did their daily jobs without using PPE. Of those workers who sprayed chemicals, only a few workers wear PPE, but they did not follow the direction of winds. The shortage offacilities such as drinking water, shower, toilets, and hand washing facilities were noticed in the firms. Therewere no warning signs at the entrance, particularly in the green house department. In the cold room department, both female and male workers had been working without the use of PPE. In many circumstances, chemical sprayers, supervisors and their assistants did not consider wind directions. The reasons mentioned were lack of awareness of how to work with chemicals, lack of PPE provision based on the importance and nature of activities and the limited supply of PPE itself made it difficult to replace when it was damaged by chemicals (Fig. 3).

The bivariate linear regression analysis showed that there was statistically significant association between the knowledge of occupational hazards and sex, age, residence, type of employment, employment duration, level ofeducation status health and safety training and orientation, and safety instructions in local language. The multivariate regression analysis identified that level of education, type of employment and work experience were factors that considerably affected the level of knowledge on occupational hazards. The level of knowledge onoccupational hazards was more likely to increase by a factor of 20.03 for respondents who had diploma and higher education than those who had no formal education [AOR: 20.03; 95% CI (16.30, 23.75)]. Respondents with > 2 years of work experience were 5.97 times more likely to be knowledgeable about workplace hazards than those who had < 2 [AOR: 5.97: 95% CI (4.22, 7.72)]. Knowledge on occupational hazards was 2.34 times more likely to increase among participants who received health and safety trainings than those who did not [AOR: 2.34; 95% CI (1.73, 3.95)] (Table 4).

In the bivariate linear regression analysis, level of education, work experience, conditions of employment, health and safety trainings, presence of safety communications/signs, like symbols and labels in the local language, observance of work rules and regulations, regular communication and supervision on health and safety, use of material safety data sheets, provision of comfortable PPE, knowledge on occupational hazards and regular use of PPE were factors significantly associated with the level of safety practice.

In the multivariate linear regression analysis, the provision of comfortable PPE, regular use and storage of PPE, presence of safety communications/signs, and knowledge on workplace hazards markedly influenced workers’ level of safety practice. The provision of PPE increased the levels of safety practice 4.59 times more likely compared to no provision of it [AOR: 4.59; 95% CI (2.34, 6.86)]. The level of safety practice was 5.15 times higher for respondents who worked where health and safety communications available than in the absence of such communications [AOR: 5.15; 95% CI (1.81, 8.49)]. The level of safety practice was 7.29 times higher for participants who had knowledge on occupational hazards than those who had not [AOR: 7.29; 95% CI (3.87, 10.73)] (Table 5).