This study utilized a descriptive cross sectional design of quantitative approaches, where the data from a population or a representative of subset is analyzed at specific point of time and the investigator measures the outcome and the exposure in the study participants at the same time (Gordis, 2009). Cross sectional study design is cheap and yet good at estimating the prevalence of the problem because data on the cause and outcome are collected at the same time [17] . This design was used because it is reliable at estimating the magnitude (compliance) and determinants of the phenomenon.

The study was carried out in Mulago National Referral hospital, department of Obstetrics and Gynecology located in Kawempe Division, approximately 5km from Kampala city center. It is the National Referral Hospital for the country and it handles all issues concerning women’s health including prenatal care, postnatal, immunization, family planning, oncology and gynecology. The hospital conducts approximately 80 deliveries per day, 2400 deliveries per month and 28,800 deliveries per year (HMIS, 2017).

The clinical services delivered in this department (labour suite) per month are offered by HCWs of different professional cadres, which include 73 midwives, 19 obstetricians 42 Senior House Officers (SHO’s), and 28 medical interns (total 163). This makes a total of 163 HCWs who work in labour suites as per monthly duty roster (Hospital, 2017). There are two units in this department that deal with deliveries these include, low risk labour suite, high-risk labour suite. The mothers after delivery are transferred to post-natal unit and of these, a bigger proportion develop puerperal sepsis (HMIS, 2017).

The study population comprised of HCWs (midwives, nurse midwives, intern nurses and midwives, senior house officers, obstetricians) who are 163 in total as per monthly duty roster (Hospital labour suite, statistics) at Mulago hospital working in the department of obstetrics and gynecology in labour suite units.

The study included HCWs (Nurses, midwives, intern nurses, nurse/midwives, intern doctors, obstetricians and senior house officers) who are involved in the management of labour.

The study excluded healthcare workers who are sick, those on leave and those busy with administrative work.

The study utilized non probability quota sampling proportionate to size using convenience sampling. A fair representation of all cadres was considered by selecting a participant from all cadres to be included in the study. The first respondent was picked on as she/he commences on her shift. No respondent declined in participating in the study if they met the inclusion criterion until the desired sample size was achieved.

The number of study subjects was estimated using Kish & Leslie (1969) formula for homogeneous population who are in single study group with proportion outcome variable.

N = z 2 p q d 2

where:

N = the required sample size.

z = the z score corresponding to 95% confidence level (1.96).

p = proportion complying with standard precautions of infection control among healthcare workers in Kenya.

q = 1 − p (the expected proportion of the study population who do not comply with standard precautions for infection control.

d = degree of precision in the study.

This formula was used to calculate N for this study, basing on the following evidence from studies carried out before.

The confidence level of the researcher in this study was at 95% (Hence z = 1.96).

Compliance among healthcare workers in Tanzania stands at 49.7% [18] .

Hence p = 0.497, the proportion, q (1 − p) will be 1 − 0.479 = 0.503 hence q = 0.503 in this study.

The degree of precision in the study was estimated at 5% (Hence d = 0.05).

Substituting in the above formula, therefore:

N = 1.96 2 × 0.497 × 0.503 0.05 2 = 384

Since the sample, size was higher than the number of the healthcare workers who are on a monthly roster of the two labour suites of Mulago hospital

The required sample size was calculated using a formula for infinite populations. Considering the finite population correction factor;

n = n 0 ⋅ N n 0 + ( N − 1 )

where n = required sample size, n₀ = 384 and N = 163, therefore by substituting in the above formula, the required sample size n = 163 respondents

N = 384 × 163 384 + ( 163 − 1 ) = 115

Hence, the sample size was 115 participants

For representativeness, Probability Proportionate to Size (PPS) was considered.

This was a non-intrusive as the healthcare workers were left to do what they usually do while practicing infection control and prevention measures in labour suite unless there is a life-threatening event that forces the observer to intervene. If this happened, then that case would be dropped, but this never happened throughout the study. During the direct observation by the principal investigator created a rapport with the study participant to pave a way to ask the few questions which are in the questionnaire indirectly in a manner that the respondent never considered this an observation of their practices. The observed data was entered on the checklist immediately to avoid recall bias. This method was adopted from a study which was done in Nigeria [1] .

Every day, at least 2 healthcare workers study participants were observed and interviewed. Data were collected between February and April 2018.

One respondent was observed at a time throughout the three procedures. The four main procedures considered included; physical examination, vaginal examinations, invasive procedures and conducting delivery. The six infection control and prevention domains were observed across these procedures. The domains have been identified as crucial for the safety of patients and HCWs by WHO, CDC and the Ugandan Ministry of Health and these are:

1) Hand hygiene;

2) Use of personal protective equipment;

3) Disinfection of reusable equipment and cleaning of delivery surfaces

4) Waste segregation of sharps

5) Waste segregation excluding sharps and,

6) The non-touch technique.

For each of these domains, compliance with Standard Precautions of Infection control (SPIC) as practiced by the healthcare workers. Each domain was scored as either “Yes” for complied or “No “not complied after the respondent has scored all. The direct observation was done without the knowledge of the study participants in order to avoid change of behavior in their SPIC practices.

Administrative clearance from Mulago hospital institutional review board was sought. Institutional consent was obtained from the unit managers to minimize on the change of behavior. Healthcare workers were identified by use of a duty roster before a shift. Since the study was unobtrusive (covert) direct observation, deferred consent was sought after the study where the principal investigator debriefed the participants about the study and the findings and asked them whether they would allow their findings be used for data analysis. In addition, those who objected this their findings were not included for data analysis.

An observation checklist was modified from CDC (2008) infection control assessment tool and a validated observation checklist of a study, which was done in Kenya, to assess compliance by practices of the HCWs on SPIC [5] .

A pilot study of 10 HCWs was conducted at Naguru Regional Referral Hospital labour ward, which has similar characteristics as the study area to test the tool for validity and reliability.

The Principal Investigator collected data in order to avoid bias by other researchers and directly observed the performance of HCWs to avoid change of behavior of HCWs during performing of routine procedures.

The questionnaire was pretested before data collection on 10 HCWs from Naguru hospital since it has similar characteristics of being also a referral facility as Mulago national referral hospital.

To improve validity/reliability of data collection on observational studies, the principal investigator through participative observation offered an assistive role to the healthcare worker performing the procedure. Interviews were conducted alongside the participative by the principal investigator. This was adopted as strategy to minimize the Wharthone’s effect.

Quality of data collected was ensured by cross checking completeness of the data checklist and the questionnaire, which was used to assess the individual and facility factors. The principal investigator also crosschecked all filled data collection tools at the end of each data collection shift.

Data collected was entered, cleaned and analyzed using SPSS version 20 to generate descriptive statistics.

Overall compliance and non-compliance was calculated by adding up all percentages from each domain then divided by the total number of domains which was 6 this method was obtained from different studies which were conducted on compliance with infection control practices [5] .

A binary score of yes (performed) and no (not performed) were used for the six domains of SPIC.

In addition, to generate descriptive statistics, categorical variables were described in form of frequencies, percentages and texts.

For objective one, data was analyzed to yield percentages of compliance of respondents and was be presented in form of a table and pie chart.

The objective two and three data were analyzed by chi square test/cross tabulation to yield frequencies, percentages and p-values for each SPIC domain. This method was utilized because the some of the data had no values hence this would not yield the Odds Ratios (OR) and Confidence Intervals (CI).

Approval to do the study was sought from the School of Health Sciences Research and Ethics Committee/institutional review board (SHREC:REF: 2017:057.) An administrative clearance was obtained from Mulago Hospital research and ethics committee. The approval letters from the School of Health Sciences REC and administrative clearance from Mulago hospital were presented to the clinical head of and of obstetrics and gynecology of Mulago national referral hospital.

Differed consent was sought after the study where the principal investigator debriefed the participants about the study and the findings and asked them whether they would allow their findings be used for data analysis. Those who objected this their findings were not included for data analysis.

The social demographic characteristics in this study were also regarded as individual factors. Data from 103 study participants were analysed. Results indicate

p-values obtained from a logistic regression. *indicates significant findings because the p-value < 0.05.

p-values obtained from a logistic regression*indicates significant findings because the p-value < 0.05.

that the majority of the study participants were females at 74%, this could have been due to the fact that majority according to professional cadre were midwives (48%). According to the Ugandan training system in midwifery education, entry requirements suggest that one should be a female in order to be enrolled in the training program. According to Kopolo et al. (2005), females are seen to be more in the clinical practice of the healthcare system and the males in administrative and policy levels.

This study demonstrated that compliance with standard precautions of infection control was generally highly associated with the above mentioned. Although the overall score was low, obstetricians in all the domains of SPIC had lower scores compared to other professional cadres for example in the hand hygiene domain where 100% of this cadre did not comply. This could have been as a result of work overload as it is indicated in Uganda the doctor patient ratio currently stands at 1:25,725 (UDHS, 2016).

It was found that majority of the medical intern (60%) were more compliant with hand hygiene compared to other professional cadres. The findings in this study were also congruent with those a study done in Jordan [19] which found out that nurses had significant higher compliance than physicians with mean = 4.91 (p = 004).

In our study, females scored higher than males in overall compliance with a statistically significance difference. This was in agreement with a study done in Jordan about factors influencing compliance with infection control precautions among 211 nurses and physicians from private and public hospitals, it was indicated that there was a significant positive correlation between compliance and individual factors such as gender where being a female was likely to be associated with compliance to SPIC. The female HCWs’ compliance with SPIC might be due to the natural proneness of female employee to obey and execute institutional rules and regulations [10] . Moreover, in Uganda women take a major role in cleaning, cooking and caretaking this could make them cautious about infection control and prevention to their family. of their family Females are usually seen performing duties that require cleanliness as part of their gender roles.

In this study work experience of 2 - 4 years was significantly (p value 0.043) associated with compliance with waste segregation of sharps. This is in contrast with pertinent studies where having more experience in clinical services was found to be significantly associated with compliance. For instance a study done in India [20] which indicated length of time was statistically significant with compliance (p = 0.014) where those who had longer stay on job had better compliance. Our study findings of a short work experience could be related to the fact that these healthcare workers are fresh graduates and inspired in applying knowledge into skills, this could also be explained by the Bloom’s taxonomy of the learning domains, where the higher the level of learning, the more the learners will tend to apply more of the psychomotor domain.

Results in this study, indicated that work in multiple places influenced compliance with hand hygiene and segregation of waste excluding sharps negatively. It was apparent that respondents working in multiple places showed lower compliance scores compared to those who did not. Work overload has also been found to influence stress and hence decrease competence among healthcare workers [21] . This could be due to the fact that, Mulago as a national referral hospital, receives an overwhelmingly huge numbers of mothers in labour yet there is inadequate staffing. In Uganda currently the doctor-patient ratio stands at 1:25,725 and nurse-patient ratio at 1:11,000 (UBOS, 2012). The doctors tend to get involved in dual practice in order for them to cope with low pay in the public health facilities such as Mulago National Referral Hospital. The results from this study agree with previous studies such as study conducted in Brazil [22] and another study in China [23] , while a study done in Italy [24] showed lower scores of 15% among healthcare workers who had a side job. This conforms to Piai-Morais et al. (2015) in his conclusion that work overload negatively affects compliance.

In this study, aseptic non-touch technique showed statistical significance with professional cadre and level of education both at Univariate and at bivariate analysis. It was also indicated that Obstetricians were 6 times more likely to comply with the asepsis non-touch technique compared to the midwives. Aseptic non-touch technique is one of the core components of SPIC, which involves a set of actions, aimed at preventing transmission of pathogenic infections to the patient (CDC, 2010). Aseptic non-touch technique vocabulary has been used interchangeably with other words which changes meaning and causes confusion which could contribute to practice variability [25] . The results from this were in congruent with the findings from a study done on two London hospitals [25] where Physicians were more compliant than the nurses in observing the non-touch technique. This could be explained by the different training systems that could have an impact on the practice outcomes especially when it comes to clinical practice [26] .

In this study, it was also found out that having had an in-service training on standard precautions of infection, control and having seen SOPs were associated to compliance with waste segregation of sharps, waste segregation excluding sharps and observing the Aseptic non-touch technique. The Odds Ratios at bivariate analysis indicated those who had seen SOPs and those who had gone through in-service training in SPIC were two times more likely to comply with these domains of standard precautions of infection control. The main purpose of standard operating procedures is to inform all healthcare workers of the correct steps to follow during the practice of SPIC this is also seen to improve knowledge on SPIC [27] . In a quasi-experimental study on the effectiveness of SOPs in prevention of needle stick injury among nurses in an Indian hospital, results showed improvement in knowledge between day 9 and 70 with post score of (22.26) on day 70 [27] . In our study, there was significant association between compliance with SPIC and having seen SOPs at Univariate, bivariate and multivariate analysis. This phenomenon is best explained by the behaviourist theory of learning where learning actually occurs when new behaviours are acquired through associations between stimuli. This means that having received training on SPIC, the SOP will be a reminder for the healthcare worker. The WHO (2016) strongly emphasizes that standard operating procedures are in place in all healthcare service delivery systems.

This proves the opportunity to train healthcare workers in SPIC and the management’s commitment on patient safety during stay in hospital with aim of preventing Hospital Acquired Infections (HAIs). Training of HCWs on SPIC is a fundamental and immediate need. This need can be identified in a study done in China [23] among 144 nurses where it was evident that less than half of the participants were knowledgeable about SPIC and 98.20% expressed interest to be trained in such measures.

It was also apparent in a recent pragmatic evaluation study in 2 London hospitals [25] where a bigger proportion of staff (37/49, 76%), thought that their practice had greatly improved and 98% of these staff considered the training had helped to standardize aseptic technique practices.

Similar findings were shown in a study done in Brazil [22] where practice of SPIC was highly influenced by availability of standard operating procedures at the hospital and continuous medical training on SPIC. In contrary, an observation study on compliance with infection control and prevention in Kenya Bedoya et al., (2017), concluded that there was a weak association between compliance and availability of all infection control supplies. In service training to all healthcare workers, despite of their professional cadres and academic qualification is vital concerning infection control and prevention in labour suites in order to prevent hospital-acquired infections to all mothers who are in labour.

The results from this study reveal lack of strategic approach for infection control and prevention in the labour suites of Mulago national referral hospital. It was shown in this study that there are no standard operating guidelines in this institution, yet there are several strategies set by the WHO for infection control and prevention in maternal health facilities for developing countries [1] . These strategies include development of an institutional infection control programme which consist of use of infection control and prevention manuals, education and training of HCWs, infection control teams and committees and audit of maternal deaths [28] .