Antimicrobial stewardship: Attitudes and practices of healthcare providers in selected health facilities in Uganda.

Though antimicrobial stewardship (AMS) programmes are the cornerstone of Uganda's national action plan (NAP) on antimicrobial resistance, there is limited evidence on AMS attitude and practices among healthcare providers in health facilities in Uganda. We determined healthcare providers' AMS attitudes, practices, and associated factors in selected health facilities in Uganda. We conducted a cross-sectional study among nurses, clinical officers, pharmacy technicians, medical officers, pharmacists, and medical specialists in 32 selected health facilities in Uganda. Data were collected once from each healthcare provider in the period from October 2019 to February 2020. Data were collected using an interview-administered questionnaire. AMS attitude and practice were analysed using descriptive statistics, where scores of AMS attitude and practices for healthcare providers were classified into high, fair, and low using a modified Blooms categorisation. Associations of AMS attitude and practice scores were determined using ordinal logistic regression. This study reported estimates of AMS attitude and practices, and odds ratios with 95% confidence intervals were reported. We adjusted for clustering at the health facility level using clustered robust standard errors. A total of 582 healthcare providers in 32 healthcare facilities were recruited into the study. More than half of the respondents (58%,340/582) had a high AMS attitude. Being a female (aOR: 0.66, 95% CI: 0.47-0.92, P < 0.016), having a bachelor's degree (aOR: 1.81, 95% CI: 1.24-2.63, P < 0.002) or master's (aOR: 2.06, 95% CI: 1.13-3.75, P < 0.018) were significant predictors of high AMS attitude. Most (46%, 261/582) healthcare providers had fair AMS practices. Healthcare providers in the western region's health facilities were less likely to have a high AMS practice (aOR: 0.52, 95% CI 0.34-0.79, P < 0.002). In this study, most healthcare providers in health facilities had a high AMS attitude and fair AMS practice.

Introduction

Antimicrobial resistance (AMR) is a global public health threat caused by the misuse of antibacterial agents in human, animal, and environmental sectors [1,2]. Antibacterial misuse involves prescribing antibacterial agents when not needed, while antibacterial overuse involves inappropriate or unnecessary taking of antibacterials [3-5]. Several international, national, and professional organisations, including the World Health Organization (WHO), have called for the establishment of antimicrobial stewardship as a strategy to promote optimal antibacterial use in the human, veterinary, and agricultural sectors in order to reduce the transmission and development of antimicrobial resistance (AMR) [6-8].

Antimicrobial stewardship (AMS) is a set of synchronised interventions that optimise antibacterial use to generate the best clinical outcome, increase patient safety, and reduce the risk of AMR development [9-11]. Together with infection prevention and control (IPC), medicine and patient safety, AMS is one of three "pillars" of an integrated strategy used in strengthening health systems [10]. Adopting AMS interventions in health facilities is critical in supporting healthcare providers with tools and systems in optimising antibacterial use, reducing the transmission and colonisation of multidrug-resistant bacteria, and lowering the incidence of antimicrobial-related adverse events [12]. Additionally, the principles of AMS are extensively applied throughout the One-health approach in optimising antibacterial use in both animal and agriculture sectors, where the emphasis is put on judicious and prudent antibacterial use to avoid the spread and development of antibacterial resistance [10,13]. Furthermore, following the approval of the Global Action Plan (GAP) on AMR, member states of the WHO committed themselves to the development and implementation of National Action Plans (NAPs) on AMR [3,14,15]. Establishing AMS programmes has been prioritised in all national action plans on AMR as a critical objective for optimising antibacterial use [10]. In developing NAPs on AMR, member states of WHO were encouraged to involve relevant stakeholders in different sectors, including institutions, health professionals, policymakers, and patients [3]. Over 117 countries have established NAPs on AMR, with varying stages of implementation of antimicrobial stewardship (AMS) programmes [3,14,15]. According to a recent systematic review, only seven African countries had NAPS on AMR, and among these, only three countries (Kenya, South Africa, and Tanzania) were implementing AMS programmes [11]. However, the same review identified that AMS activities were implemented in countries with neither NAPs nor AMS programmes [11]. On the other hand, previous studies have shown that the adoption of AMS activities is dependent also on the healthcare providers’ attitudes and practices, which were affected either by the top-down approach of government policy implementation or bottom-up participation of healthcare providers in policy development [8,16].

Previous studies on AMS showed that healthcare providers’ attitudes and practices on AMS varied significantly throughout most countries like Nigeria, Zambia, and Ethiopia [17]. In most health facilities, healthcare providers had a casual and lax attitude towards AMS implementation following its introduction through a top-bottom approach to policy implementation [17]. In addition, despite having good Knowledge about AMS, most healthcare providers did not agree that antibacterials were misused or that AMR was a significant problem in their institutions [17]. Several studies show that the top-down approach has limited the implementation of AMS programmes and activities to only specialised hospitals, thus disregarding community hospitals whose major labour force are allied healthcare professionals [16-18]. As a result, this has limited the application of previous findings on antimicrobial stewardship attitudes and practices to only specialised health care facilities and a particular group of healthcare providers. The exclusion of community health facilities in previous studies in low-and-middle-income countries (LMICs) has limited the generalizability of their findings on AMS attitudes and practices of healthcare providers in their countries [19].

Several one-health initiatives have trained healthcare providers in health facilities and communities on implementing antimicrobial stewardship in both the human and animal sectors in Uganda [13]. Additionally, when Uganda drafted its National Action Plan (NAP) on AMR for 2018 to 2023, it placed a greater emphasis on a bottom-up approach that included healthcare providers in regional referral centres, general hospitals, and private not-for-profit (PNFP) organisations, as well as other stakeholders to promote AMS [20]. Despite the Ministry of Health’s continued engagement with heads of health institutions in strengthening existing medicine and therapeutic committees and antimicrobial stewardship programmes, the attitudes and practices of healthcare providers towards AMS remain unknown in Uganda [13,20]. Therefore, the current study investigated healthcare providers’ attitudes and practices towards AMS and associated factors in Uganda’s regional referral hospitals, general hospitals, and private-not-for-profit (PNFPs) health facilities.

Materials and methods

Study design and setting

We conducted a cross-sectional study from October 2019 to February 2020 among healthcare providers at regional referral hospitals, general hospitals, and PNFPs. Uganda’s health system is integrated with about 6,937 health facilities, where 45% are public-owned, 40% are private for-profit (PFP), and 15% are private-not-for-profit (PNFPs) [21,22]. The public health system is hierarchical, referral-based, and provides free health services at all levels of delivery [22,23]. The composition of public health facilities in the country includes; two national referral hospitals, 16 regional referral hospitals, 47 general hospitals, 166 level IV health centres, 962 level III health centres, and 1321 level II health centres and 1558 clinics [21,23,24]. The two National referral hospitals are urban teaching hospitals with a bed capacity ranging from 600 to 1500. Regional referrals hospitals are teaching hospitals located in urban centres with a bed capacity ranging from 250 to 600 beds, while general hospitals are community-based, all have 100 beds. There are 1009 health facilities, four tertiary hospitals, followed by 40 general hospitals that serve community settings and 955 health centres [22,24]. Due to government funding, they offer cheaper, subsidised services than PFP [21,23]. There are 2976 health facilities in the PFP healthcare systems of Uganda [21].

The study was conducted at selected regional referral, general hospitals and tertiary PNFP hospitals in all regions of Uganda. The hospitals above were selected because their healthcare providers and those in the health centres and communities had received training on antimicrobial stewardship through several one-health initiatives [13]. Additionally, the government is strengthening health systems by operationalising medicines and therapeutic committees to strengthen supply chain management, antimicrobial stewardship programmes, and pharmacovigilance [20,25].

Study population

The study population included hospital directors, nursing officers, heads of department (Medicine, Paediatrics, Surgery and Obstetrics and Gynaecology), medical officers, clinical officers, pharmacy technicians, pharmacists, and medical specialists. In this study, pharmacy technicians, referred to as dispensers had a pharmacy diploma while pharmacists had a bachelor’s degree. Clinical officers, referred to community health officers, had a diploma in clinical medicine, while medical officers and medical specialists had a bachelor’s degree and a master’s degree in medicine, respectively. The study included only permanent staff who had worked for more than two years. Part-time staff, medical residents, interns and those who had worked for only one year after transferring to the health facility were not included.

Sample size determination

The required sample size was determined using a single population proportion formulae [26]. We took the proportion to be 50% (P = 0.5) to have a maximum sample size possible with the formulae since there were no previous studies on AMS attitudes and practices. Using a population of 42500 healthcare providers with a 5% margin of error, we obtained a sample size of 381. We adjusted for clustering by multiplying the sample size (381) obtained by 1.5 and adjusted for a non-response rate of 35.4%, thus generating a sample size of 768 health providers. The study targeted eight regional referrals, 32 general hospitals and three tertiary PNFP to achieve the required sample size. In each facility, we targeted participation of 17 to 24 health providers to reach the estimated sample size.

Sampling procedure

The sampling frame consisted of 16 regional referrals, 47 general hospitals and four tertiary PNFPs. The health facilities were selected because their healthcare providers had previously received AMS training from numerous One-health initiatives.

We selected eight regional referral facilities out of 16 (50%) using simple random sampling (lottery method). We then selected 32 general hospitals out of 47 (68.1%) using simple random sampling (lottery technique). When the selection procedure was completed, four general hospitals were selected for each of the eight regional referral hospitals in each of the country”s regions. Lastly, we randomly selected three out of four (75%) tertiary PNFPs hospitals. We sought administrative clearance (consent to participate) of all the selected health facilities (eight regional referrals, 32 general hospitals and three tertiary PNFPs). The study received administrative clearance from eight selected regional referral hospitals, only 21 out of 32 general hospitals, and three tertiary PNFPs.

We selected healthcare providers using a proportionate number to size. We targeted 224 nurses, 192 clinical officers, 32 pharmacy technicians, 194 medical officers, 32 pharmacists, 64 medical experts, and 32 laboratory technicians out of the needed 768 healthcare providers. We computed the number of different professionals to be selected from each facility by dividing the number of people in a specific profession by the total number of health professionals to obtain the fraction of that profession at the facility. This fraction was then multiplied by the total number of health professionals to be sampled from the health facility. The different numbers of healthcare providers per health facility selected in the study are shown (S1 Fig supplementary information).

Survey tool development

We conducted an extensive literature review with keywords related to antimicrobial stewardship, attitude and practices to generate a pool of questions. The questionnaire items on AMS attitude included healthcare providers’ familiarity with AMS, the effectiveness of AMS in improving patient outcomes, patient safety, and reducing the spread of antibacterial resistance [17,27]. AMS practice items included; adherence to standard treatment guidelines, culture and susceptibility testing, avoidance of excessive use of broad-spectrum antibacterials, and surgical antibacterial prophylaxis [17,28-31]. The pharmacologist prepared the initial version of the instrument in English since it targeted only healthcare practitioners whose formal language of practice is English. We invited specialists in public health (1), epidemiology (1), microbiology (1), pharmacy (1), and pharmacology (1) to review and modify the instrument to improve the clarity of each item’s questions, ease of comprehension, and style and structure of the questionnaire.

Furthermore, specialists agreed on the AMS attitudes and practices questions’ readability, clarity, and comprehensiveness. We pilot-tested the questionnaire in four hospitals with 20 healthcare practitioners (doctors (6), nurses (4), allied health workers (8), and pharmacists (2). The respondents gave feedback on questions that needed reformulation, rewording, as well as those that were difficult to understand. We tested the reliability of the piloted tool by conducting an alpha Cronbach’s coefficient, where AMS attitude was 0.9268 and 0.762 for AMS practice. After approval from the experts, the final version of the questionnaire contained arranged attitude and practice questions according to the respondents’ breadth and depth of understanding of their particular hierarchies (Bloom, 1956) (S1 Appendix supplementary material).

Variables

The outcome variables were attitude and practices on AMS. AMS attitude and practice scores were generated as the sum of points in each of the 12 questions. The questions on AMS attitude were Likert type, and responses scores were coded from 1 to 5 (strongly disagree = 1, disagree = 2, neither agree nor disagree = 3, agree = 4, strongly agree = 5), giving a possible minimum of 12 and a possible maximum of 60 points in all 12 questions. The questions on AMS practices required "yes" (coded as 1) or "no" (coded as 0) responses and thus had a minimum possible score of 0 and a maximum possible score of 12. The study graded healthcare providers’ AMS attitude and practice scores using modified Bloom’s categorisation [32]. According to this study, AMS attitude or practice scores had a "high" score if they ranged between 80 and 100% (47–60) points for attitude and over ten points for practice, "fair " if the score was between 50 and 79% (30–46 points for AMS attitude, and 6 to 9 points for AMS practices), and "low" if the score was less than 50% (30 points for attitude and less than 6 points for AMS practice). Independent variables included both social-demographic (sex, age, years of experience, level of academic training, healthcare provider’s profession) and hospital characteristics such as the type of health facility (general, regional referral, and private-not-for-profit), nature of health facility (teaching and non-teaching hospitals) and bed capacity.

Data collection

Data were collected using an interviewer-administered questionnaire. Before data collection, research assistants, comprising medical officers, pharmacists, nurses, and hospital biostatisticians, were trained on the questionnaire. The survey questionnaire had sections on: (i) hospital characteristics; (ii) respondents’ socio-demographic characteristics; (iii) AMS attitude; (iv) AMS practice.

To recruit study respondents, research assistants used phone calls, text SMS, emails and letters to invite all selected healthcare providers from the Departments of Medicine, Paediatrics, Surgery, Obstetrics and Gynaecology, Private, Outpatients, Pharmacy, and Laboratory for the interview. We used a list of health workers obtained from the medical director’s office or heads of departments. Research assistants made reminder phone calls to the potential study participants to increase participation. We informed every recruited respondent that their participation was voluntary. After accepting to participate, the research assistant provided a brief introduction of the study, objective, and procedures and informed the respondent of anonymity, confidentiality and all declarations. Responding to the questionnaire took 25 to 30 minutes. The respondents were compensated for their time and transport. We collected data in each health facility for two weeks from November 2019 and February 2020.

Data processing and management

The research assistant evaluated every questionnaire for accuracy and completeness at the end of each day’s data collection. During fieldwork and data cleaning, we performed a thorough case analysis to detect missing data on variables in the questionnaires. We dropped any questionnaires containing significant missing data on study variables during the data cleaning process. We utilised EpiDATA manager to conduct double data entry and validation during which data collection tools were entered twice by different data entrants, which we reconciled to detect any differences or discrepancies. Those variables which diverged from each other were thoroughly checked against the original questionnaires and harmonised accordingly. We performed the data validation during entry until the original and entered files were similar to each other.

Data analysis

All data collected was analysed using STATA 15.1 (Stata Corp, Texas, USA). The study summarised categorical variables using proportions, and it further described continuous variables using means and standard deviations or medians and interquartile ranges (IQR). Socio-demographic variables associated with the AMS attitude and practice scores were determined using ordinal logistic regression in bivariable and multivariable analysis. After testing the association between AMS attitudes or practices with social demographics, two variables (nature of teaching health facility and bed capacity) had a variance inflation factor (VIF) greater than 10, which indicated the presence of multicollinearity. The study chose the type of teaching health institution over bed capacity because it had a lower Bayesian information criterion (BIC) value.

We included all variables with a p-value less than 0.2 at bivariable analysis in the multivariable analysis. Two independent variables, the region of Uganda and the health facility department, violated the proportional odds assumption. There was no significant difference between non-proportional and proportional odds models using BIC. So all ordinal logistic regression used proportional odds models. We included age and sex as universal confounders even when they did not reach the 0.2 significance criterion in the multivariable model using the backward selection technique, along with all variables with a p-value less than 0.2 in the bivariable model. The dependent variable had three categories; "low," "fair," and "high" attitude or practice scores of AMS. Independent variables were assessed for statistical interactions and confounding. In this study, variables with p-values less than 0.05 were considered statistically significant in the final model, where sex and age were used as universal confounders. Associations of AMS attitudes and practices were presented using odds ratios and their corresponding 95% confidence intervals. The research used clustered robust standard errors to account for health facility clustering.

Ethical considerations

The protocol received ethical clearance from the Makerere University School of Biomedical Sciences Higher Degree Research and Ethics Committee (reference number SBH-HDREC-624). The study got further ethical approval from the Uganda National Council of Science and Technology (UNCST) and gave ethical approval (reference number HS339ES). Heads of health facilities granted the study protocol administrative clearance, permitting the principal investigator to conduct the study among healthcare providers in all participating health facilities. Before responding to the questionnaire, written informed consent was obtained from all targeted respondents. We kept all the questionnaires collected from the survey in lockable lockers for confidentiality. All information about the healthcare providers was de-identified to ensure anonymity.

Results

Socio-demographic characteristics of study respondents

Of the 768 potential study respondents contacted for enrolment from 32 health facilities, 582 completed the study questionnaire (76%, 582/768). More than half of the study respondents were females (57%, 333/582). The overall median age of the respondents was 38 (IQR, 34–43) years. Most of the respondents (42%, 246/582) were between 30 and 39 years old. Over half (56%, 327/582) of the healthcare providers had a diploma level of academic training. Most healthcare providers (44%, 258/582) had worked for more than ten years (Table 1).

Table 1

Characteristics of study respondents (N = 582).

Description	Frequency (N = 582)	Percentage (%)
Sex
Females	333	57.2
Males	249	42.8
Age (years)
20–29	96	16.5
30–39	246	42.3
40–49	171	29.4
50+	69	11.9
Level of academic training
Diploma	327	56.2
Degree	191	32.8
Masters	64	11
Years of experience
Less than five years	184	31.6
5 < 9	140	24.1
10+	258	44.3
Healthcare providers
Nurses	199	34.2
Pharmacy Technicians (PTs)	30	5.2
Clinical Officers (COs)	136	23.4
Medical Officers (MOs)	121	20.8
Pharmacists (P)	24	4.1
Medical specialist (MS)	50	8.6
Laboratory technicians (LTs)	22	3.8

Antimicrobial stewardship attitudes of healthcare providers in health facilities in Uganda

More than half of the healthcare providers enrolled in this study had a high AMS attitude (58%, 340/582). Pharmacists had the highest mean AMS attitude scores compared to all healthcare providers. Nurses had the least mean AMS attitude scores. Healthcare providers agreed that implementation of AMS strategies in health facilities minimises the risk of antibacterial resistance development (87%, 507/582), decrease patient length of stay (85%, 496/582), improve patient outcomes (82%, 492/582) and increase appropriate antibacterial use (81%, 474/582) (Table 2).

Table 2

Attitudes of healthcare providers on antimicrobial stewardship (AMS) in health facilities in Uganda (N = 582).

	Healthcare providers in selected health facilities (N = 582)
	Nurses	PT	CO	MO	P	MS	LT	Total
	(n = 199)	(n = 30)	(n = 136)	(n = 121)	(n = 24)	(n = 24)	(n = 22)	580
Antimicrobial stewardship (AMS) attitudes	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(100)	P-Value
I know what AMS means	111 (55.8)	20 (66.7)	78 (57.3)	84 (69.4)	24 (100.0)	33 (66.0)	14 (63.7)	364 (62.5)	0.001
I am familiar with AMS goals	74 (37.2)	13 (43.3)	54 (39.7)	55 (45.5)	16 (66.6)	29 (58.0)	7 (31.8)	248 (42.6)	0.001
AMS is essential in this health facility	153 (76.9)	26 (86.7)	104 (76.5)	102 (84.3)	22 (91.7)	41 (82.0)	13 (59.1)	461 (79.2)	0.004
AMS involves appropriate selection of antibacterials	137 (68.8)	23 (76.7)	104 (76.4)	97 (80.2)	22 (91.6)	37 (74.0)	12 (54.6)	432 (74.0)	0.001
AMS involves optimal antibacterial administration	144 (72.4)	22 (73.3)	101 (74.3)	93 (76.9)	22 (91.6)	37 (74.0)	15 (68.2)	434 (75.0)	0.023
AMS interventions can improve patient outcomes	170 (85.4)	28 (93.3)	114 (83.9)	104 (86.0)	23 (95.8)	44 (88.0)	19 (86.3)	502 (86.0)	0.094
AMS strategies can reduce the problem of antimicrobial resistance	172 (86.4)	28 (93.4)	117 (86.0)	105 (86.7)	24 (100.0)	42 (84.0)	19 (86.4)	507 (87.1)	0.117
AMS can reduce the length of hospital stay	168 (84.5)	28 (93.3)	112 (82.4)	102 (84.3)	24 (100.0)	44 (88.0)	18 (81.8)	496 (85.2)	0.417
AMS practices can increase appropriate antibacterial use	147 (73.8)	27 (90.0)	114 (83.8)	100 (82.6)	24 (100)	44 (88.0)	18 (81.9)	474 (81.4)	0.001
AMS strategies can decrease the incidence of Clostridium difficile rates	139 (69.9)	20 (66.7)	87 (63.9)	89 (73.5)	22 (91.7)	32 (64.0)	16 (72.8)	405 (69.6)	0.289
Source of information on AMS practices.	61 (30.7)	6 (20.0)	38 (27.9)	30 (24.8)	5 (20.8)	13(26.0)	5 (27.1)	158 (27.1)	0.943
Additional staff education on AMS is needed	171 (85.9)	28 (93.3)	116 (85.3)	103 (85.1)	24 (100)	45 (90.0)	18 (81.8)	505	0.057
AMS attitude scores reported as mean and standard deviation (SD) in each profession group	44.5±11.4	47.8±9.2	45.3±11	47.2 ±10.9	52.7±4	47.5±11	43.9±12.6	46±11

aPT: Pharmacy technician, CO: Clinical officer, MO: Medical officer, P: Pharmacist, MS: Medical Specialists LT: Laboratory technician.

bA Likert scale rated from one (strongly disagree) to 5 (strongly agree) and *show significant difference at P < 0.05.

Factors associated with antimicrobial stewardship attitude among healthcare providers in health facilities in Uganda

In a bivariable analysis, AMS attitude amongst healthcare providers was significantly associated with level of academic training (P = 0.002), hospital department (P = 0.006), sex (P = 0.005), and geographical region (P = 0.001). After controlling for education and region, the multivariable logistic regression model showed that females (AOR: 0.66, 95% Cl: 0.47–0.92) were significantly less likely to have high AMS attitude scores than males after controlling for education and the region. Healthcare providers with a bachelor’s degree (AOR: 1.81, 95% Cl: 1.24–2.63) were 1.8 times significantly more likely to have high AMS scores than those with diplomas. Similarly, healthcare providers with a master’s degree (AOR: 2.06: 95% Cl: 1.24–2.63) were 2.1 times significantly more likely to have high AMS attitude scores than those with diplomas (Table 3).

Table 3

Predictors of antimicrobial stewardship (AMS) attitudes amongst healthcare providers in health facilities in Uganda (N = 582).

	Low score	Fair score	High scores	COR	AOR	P-value
	(n = 52)	(n = 190)	(n = 340)	(95% CI)	(95% CI)
	n(%)	n(%)	n(%)
Age (years)
20–29	13(25)	29(15.3)	54(15.9)	1
30–39	16(30.8)	77(40.5)	153(45)	1.48 (0.91–2.39)	1.46 (0.90–2.37)	0.121
40–49	18(34.6)	62(32.6)	91(26.8)	1.06 (0.64–1.77)	1.05(0.63–1.74)	0.862
50+	5(9.6)	22(11.6)	42(12.4)	1.59 (0.83–3.02)	1.57(0.83–2.99)	0.168
Sex
Male	23(44.2)	97(51.1)	213(62.6)	1	1
Female	29(55.8)	93(48.9)	127(37.4)	0.65 (0.46–0.91)	0.66 (0.47–0.92)	*0.016
Level of academic training
Diploma	36(69.2)	124(65.3)	167(49.1)	1	1
Degree	13(25)	51(26.8)	127(37.4)	1.77 (1.21–2.58)	1.81 (1.24–2.63)	*0.002
Masters and above	3(5.8)	15(7.9)	46(13.5)	1.96 (1.05–3.65)	2.06 (1.13–3.75)	*0.018

COR: Crude Odds Ratio, AOR: Adjusted Odds Ratio, CI: Confidence Interval.

*show significant difference at p < 0.05.

AMS practices among healthcare providers in health facilities in Uganda

Most respondents (47%, 261/582) had a fair AMS practice score in this study. Medical officers had the highest mean AMS practices, while pharmacists had the least mean AMS practice scores.

The most-reported AMS practices implemented in health facilities included; documenting antibacterial use (90%, 501/560); using standard treatment guidelines to initiate effective antibacterial treatment (78%, 438/560); and complying with culture and susceptibility results (76%, 425/560) (Table 4).

Table 4

Practices of healthcare providers on antimicrobial stewardship in health facilities in Uganda.

	Healthcare providers in selected health facilities (N = 582)
	Nurse	PT	CO	MO	P	MS	LT	Total	P-Value
	(n = 199)	(n = 30)	(n = 136)	(n = 121)	(n = 24)	(n = 24)	(n = 22)	580
Antimicrobial stewardship (AMS) Practices	(%)	(%)	(%)	(%)	(%)	(%)	(%)	(%)
Use of standard treatment guidelines	144 (72.4)	23 (76.7)	114 (83.8)	104 (86.0)	16 (66.7)	37 (74.0)	18 (81.8)	456 (78.4)	*0.037
Avoid unnecessary broad spectrum antibacterial use	132 (66.3)	14 (46.7)	85 (62.5)	81 (66.9)	5 (20.8)	29 (58.0)	8 (36.4)	354 (60.8)	*<0.001
Documenting antibacterial use in patient care	181 (91.0)	27 (90.0)	118 (86.8)	114 (94.2)	17 (70.8)	44 (88.0)	20 (90.9)	521 (89.5)	*0.036
Pre-surgical single-dose antibacterial administration	94 (47.2)	14 (46.7)	58 (42.6)	56 (46.3)	9 (37.5)	18 (36.0)	5 (22.7)	254 (43.6)	0.319
Complying with culture and susceptibility results	150 (75.4)	21 (70.0)	98 (72.1)	93 (76.9)	18 (75.0)	45 (90.0)	9 (40.9)	434 (74.6)	*0.002
Antimicrobial prescription audit and review	146 (73.4)	23 (76.7)	106 (77.9)	94 (77.7)	17 (70.8)	34 (68)	12 (54.5)	432 (74.2)	0.275
Antibacterial time-out	141 (70.9)	19 (63.3)	94 (69.1)	85 (70.2)	16 (66.7)	34 (68.0)	9 (40.9)	398 (68.4)	0.182
Patient education on antibacterial use	134 (67.3)	21 (70.0)	101 (74.3)	89 (73.6)	18 (75.0)	32 (64.0)	14 (63.6)	409 (70.3)	0.657
Existence of antibacterial use best practices	107 (53.8)	14 (46.7)	89 (65.4)	78 (64.5)	14 (58.3)	25 (50.0)	10 (45.5)	337 (57.9)	0.09
Assessment of antibacterial use (quality and quantity)	86 (43.2)	13 (43.3)	55 (40.4)	44 (36.4)	10 (41.7)	10 (20.0)	7 (31.8)	225 (38.7)	0.111
Measurement of antibacterial use outcomes	111 (55.8)	16 (53.3)	94 (69.1)	75 (62.0)	12 (50.0)	25 (50.0)	14 (63.6)	347 (59.6)	0.121
Use of hospital antibacterial audit data	88 (44.2)	8 (26.7)	58 (42.6)	54 (44.6)	6 (25.0)	23 (46.0)	8 (36.4)	245 (42.1)	0.31
AMS practice scores reported as means and standard deviation (SD)) in each of the professional groups	7.6 ± 3	7.1±2.8	7.9±2.9	8.0 ± 2.6	6.6±3.0	7.1±2.7	6.1± 3.3	7.6 ± 2.9

PT: Pharmacy technician, CO: Clinical officer, MO: Medical officer, P: Pharmacist, MS: Medical specialist, LT: Laboratory technician, SD: Standard deviation.

*shows a significant difference at p < 0.05.

Factors associated with antimicrobial stewardship practices among healthcare providers in health facilities in Uganda

In bivariate analysis, AMS practice scores were significantly associated only with the region (P = 0.003). In the multivariable model, after having adjusted for confounders, AMS practice scores of healthcare providers in the western region (AOR: 0.52, 95% CI: 0.34–0.79) were significantly lower than those in the central region (Table 5).

Table 5

Predictors of antimicrobial stewardship practices amongst healthcare providers in health facilities in Uganda (N = 582).

	Low score	Fair score	High scores	COR	AOR (95% CI)	P-value
	(n = 133)	(n = 261)	(n = 166)	(95% CI)
	n(%)	n(%)	n(%)
Age
30–39	20(21.7)	46 (50)	26(28.3)	1	1
20–29	52(22.4)	115(49.6)	65(28)	0.99 (0.62–1.6)	1.02 (0.66–1.61)	0.901
40–49	45(26.9)	69(41.3)	53(31.7)	0.94 (0.64–1.38)	0.95 (0.65–1.39)	0.785
50+	16(23.3)	31(44.9)	22(31.9)	1.11 (0.67–1.87)	1.15 (0.69–1.91)	0.593
Sex
Male	83(26.1)	143(45)	92(28.9)	1	1
Female	50(20.7)	118(48.8)	74(30.6)	1.13 (0.75–1.69)	1.18(0.85–1.62)	0.313
Region of Uganda
Central	37(20.2)	91(49.7)	55(30.1)	1	1
North	10(15.6)	29(45.3)	25(39.1)	1.52 (0.89–2.60)	1.48(0.86–2.53)	0.153
East	35(20.3)	80(46.5)	57(33.1)	1.13(0.76–1.68)	1.1 (0.75–1.64)	0.615
West	51(36.2)	61(43.3)	29(20.6)	0.54(0.35–0.82)	0.52 (0.34–0.79)	*0.002

COR: Crude Odds Ratio, AOR: Adjusted Odds Ratio, CI: Confidence Interval.

*shows a significant difference at p < 0.05.

Discussion

Healthcare providers are more likely to their attitudes and practices if they are involved in the policymaking process and agree with the proposed changes [33]. The Ministry of Health used a bottom-up strategy, where its actively engaged healthcare providers in operationalising the implementation of AMS programmes and strengthening Medicines and Therapeutics committees in health facilities [20]. Despite the continued engagement with health facilities, healthcare providers’ attitudes and practices and associated factors towards AMS have remained unknown in all four Ugandan regions. This study uses data from an interviewer-administered questionnaire among 582 healthcare providers in 32 health facilities from October 2019 to February 2020 to explore attitudes and practices concerning antimicrobial stewardship (AMS) in selected health facilities in Uganda.

More than half (58%) of the respondents had high AMS attitude scores in our study. This finding contrasts with a previous study on AMS attitude conducted in Ethiopia, where 16% of healthcare providers had a high AMS attitude [17]. The variations in AMS attitudes between the two studies could be because of differences in participating health facilities as well as the bottom-up approach of the Ministry of Health involving health facility leaders in the strengthening or operationalisation of AMS programmes and medicine and therapeutics committees in public health facilities and PNFPs. As shown in our study, the bottom-up strategy improved commitment of healthcare providers leading to adoption of interventions that may induce behaviour change. However, a previous study in Ethiopia showed that low AMS could have arisen due to implementing restrictive AMS strategies, which potentially affected the attitude of healthcare providers [34]. While the Ministry of Health in Uganda is in the implementation stages of the NAP on AMR, which places AMS as a critical priority, the high AMS attitude in our study suggests that health facilities could have adopted AMS programmes with strategies that improve healthcare providers’ attitudes prior to the Ministry of Health policy intervention.

AMS education covers many subjects, including proper antimicrobial selection and prescription, optimising dosages and duration, and minimising toxicity and side effects to improve clinical, economic, and microbiological results[35]. As a result, prior research has emphasised the need of employing a multidisciplinary team of highly qualified pharmacists and infectious disease specialists to lead AMS programmes [36]. Our study agrees with previous findings where those with high academic training, like having a bachelor’s or master’s degree, had significantly higher AMS scores than those with diploma training. After controlling for education, females had lower AMS attitude scores than males. This demonstrates that females (56%) may have comprised a significant proportion of diploma holders who could not obtain AMS training before and during practice experience. The low AMS attitude scores of diploma holders and females could affect the implementation of AMS programmes in terms of comprehensiveness, quality, and adoption to lower community health facilities [37]. These findings suggest the need for curricula on AMS service training for all diploma holders to harmonise their attitude with those of higher qualification regarding AMS to strengthen the multidisciplinary healthcare provider capacity to perform strategies of AMS.

In our study, a third of the respondents reported high AMS practices scores. Our study finding contrasts that of a previous study in Ethiopia, which found that over 70% of healthcare providers have high AMS practices [17]. Despite the high AMS attitude reported in this study, most healthcare providers reported fair AMS practices. However, medical officers had a high mean AMS practice score compared to other healthcare providers. This finding could be an indicator of challenges in implementing the AMS programmes. The absence of national AMS guidelines for health facilities, non-functional microbiology laboratories, and many low-level healthcare cadres employed in health facilities may contribute to this fair AMS practice reported in this study [38].

In this study, reported AMS practices were significantly associated with the geographic region of Uganda. Healthcare providers in the western part of the country were less likely to report high AMS practices than other regions. A previous study conducted in the Western region of Uganda reported a lack of AMS programmes and the need to strengthen infection control practices in Western Uganda’s health facilities [39]. This lack of AMS programmes in health facilities in Western Uganda may explain our study’s observed finding. There has been a significant improvement in healthcare infrastructure in the country, where the government has constructed, renovated, and upgraded many health facilities. In addition, more healthcare personnel have been recruited, hence improving the staffing levels to over 70% in most health facilities [40]. However, there is a need for national guidelines on AMS programmes and specific AMS training for healthcare providers in health facilities.

Our finding on the high AMS attitude of pharmacists agrees with previous studies, which have demonstrated that pharmacists have a high positive attitude towards AMS [29,41,42]. However, our study found that pharmacists had a low AMS practice [29]. This finding is similar to a previous study conducted in Zambia, where community pharmacists had low AMS practices concerning AMS [29]. Much as community pharmacists in the Zambian study had a challenge of dispensing antibacterials without prescriptions, lack of providing complete counselling information to patients, pharmacists in our study were less likely to use standard treatment guidelines, avoid unnecessary use of broad-spectrum antibacterials and measure the quality and quantity of antibacterial use in their health facilities. The high AMS attitude of pharmacists in our study could be from previous training on AMS though the low AMS practices could be arising from the limited mandate pharmacist could have as professionals in patient care decision making. Unlike previous studies showing the changing role of a pharmacist inpatient care under antimicrobials stewardship, in Uganda, they are still confined to their traditional function of providing advice on proper antimicrobial utilisation and creating awareness campaigns targeting other healthcare providers about the appropriateness of antimicrobial prescribing and the use of standard treatment guidelines [43,44]. There is a need for policy intervention through the Ministry of Health to strengthen AMS programmes to expand the pharmacist’s role under a multidisciplinary team, as reported by several studies.

The study’s limitations could be attributed to social desirability, which could have arisen from respondent’s responses to different interviewers. We minimised this measurement bias by piloting the questionnaire to minimise ambiguity in questions, rephrasing and rewording the questions. Using an interviewer-administered questionnaire minimised the social desirability effect. However, interviewer bias in this study was minimised by using data collectors/ interviewers from the same hospital unit. Part-time and intern healthcare practitioners could not be included in the study since they were not on the permanent employees’ lists, even though they prescribed antibacterials. We could have missed responses from this group of persons. The study had a non-response of about 24% of the sampled health providers and a lack of administrative clearance from some facilities, which could have created selection bias. The questionnaire assessed AMS practice using "yes’ or ’no’ responses, which may have over or underestimated AMS practice among healthcare providers. Our research could not determine the cause and effect relationship of whether high AMS attitude scores also contributed to fair AMS practices. There is a possibility that our findings or conclusions are not generalisable to non-Ugandan or non-East African situations. The study used tools that had been pilot-tested whose reliability and validity was known before data collection, and this potentially reduced the likelihood of under or overestimating AMS practices. Furthermore, the high Cronbach alpha (0.7) indicated the test items’ reliability and internal consistency in the tool. The inclusion of healthcare providers of various levels of training and profession from all the regions of Uganda increased the representativeness of this study’s findings.

Conclusion

In this study, most healthcare providers reported a high AMS attitude and fair AMS practices scores. The Ministry of Health should support and regularly monitor the countrywide implementation of AMS programmes by educating all public hospital healthcare providers, since our study found a significant association between AMS attitude and practices with education levels and geographic location. There is a need for more studies to assess whether adopted AMS programmes exist in these health facilities and the characteristics and challenges of implementing AMS strategies on optimising antibacterial use.

STROBE statement checklist of items included in reports of cross-sectional studies.

(DOCX)

Click here for additional data file.

Flow diagram of the selected health facilities that participated in this study.

(TIF)

Click here for additional data file.

Questionnaire for sub-study II on antimicrobial stewardship attitudes and practices of healthcare providers in selected health facilities in Uganda.

(DOCX)

Click here for additional data file.

Supporting data for the manuscript.

(DOCX)

Click here for additional data file.

23 Jul 2021

PONE-D-21-10939

Antimicrobial stewardship: Attitudes and practices of healthcare providers in selected health facilities in Uganda

PLOS ONE

Dear Dr. Kimbowa,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Sep 06 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Elena Ambrosino

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include additional information regarding the survey or questionnaire used in the study and ensure that you have provided sufficient details that others could replicate the analyses. For instance, if you developed a questionnaire as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information. Moreover, please include more details on how the questionnaire was pre-tested, and whether it was validated.

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

4. Your ethics statement should only appear in the Methods section of your manuscript. If your ethics statement is written in any section besides the Methods, please delete it from any other section.

5. We note you have included a table to which you do not refer in the text of your manuscript. Please ensure that you refer to Table 5 in your text; if accepted, production will need this reference to link the reader to the Table.

6. Please upload a copy of Supporting Information S5 which you refer to in your text on page 292.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: No

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors present their findings on an important research topic – attitudes and practices around antimicrobial resistance in Uganda. This is an important study to help inform best practices for antimicrobial stewardship in Africa. However, I think that prior to publishing this work, additional modifications should be made to clarify the methods and results, which should be highlighted further in the discussion. I think the conclusions should also be reframed and toned down.

Study design: Cross-sectional, appropriate to the research question.

Title: Appropriate, though it should be adjusted to note the study is about attitudes in children under 5 if that is the case.

Abstract: Well-written and appropriate.

Introduction: Generally well-written and appropriate to the research material.

Lines 41-44: Can you please specify what information is lacking? Have there ever been surveys of AMS attitudes and practices from sub-Saharan African countries? If so, what were the results? If not, please note this, and also note the other specific gaps in knowledge.

Methods:

Line 53: Consider replacing the word ‘nationwide’ with another word – this study was not conducted at every hospital in Uganda. Specify in the Methods or Results section what proportion of each type of hospital participated in the study.

Lines 58-66: How was the recruitment/invitation to participate carried out? Was any compensation/incentive provided?

Lines 59-62: Why were resident/MMed and other trainee physicians and midwives not included? In some hospitals, I imagine these are the people who may often prescribe or recommend antibiotics to patients.

Lines 69-71: Please describe the sample size method in more detail – referencing the WHO method does not make it clear how the appropriate number of hospitals and number of participants was calculated.

Lines 73-74: Please add the proportion of the total facility type represented by the facilities sampled for this study (e.g. 8/16 RRH = 50%).

Lines 89-90: How was the random sampling performed (e.g. what random selection algorithm was used, and how was it implemented)?

Lines 93-95: Please be more specific about which questionnaire items were adopted from the literature, which have already been validated, and which were created de novo.

Line 110: Why did the authors choose a modified Bloom’s categorization? Has this been validated for this type of research question?

Lines 111-113: I recommend reconsidering the labels of “good” and “bad” for AMS score. These labels imply judgement about attitudes and practices. I recommend changing the labels to “high”, “average”, and “low” or something similar. Labeling people and their attitudes towards AMS as good and bad could have unintended negative consequences.

Line 114: I don’t think gender was assessed, only sex. Is this correct? If so, I recommend changing this to sex.

Lines 133-134: How were discrepancies in doubly-entered data resolved?

Ethics lines 312-313: Why didn’t all research participants give written informed consent to participate? I see a statement at the end of the manuscript that ‘many’ of the participants gave written informed consent, but not all.

I see the Appendix S3 notes the topic of the research is "Antimicrobial stewardship practices and quality of antibacterials use in children under-five in Health facilities of Uganda." Was this study specifically addressing antibacterial use in children? Or all patients?

Results:

Lines 159-160: Were the non-respondents different to the respondents in any way (e.g. professional cadre, sex, age, etc)?

Line 171 and throughout the manuscript: Please see my note about labeling AMS scores as ‘good’ or ‘bad’. The same is true for ‘poor’. I think this should be changed to ‘low’.

Lines 185-187: Please indicate directionality of the association, e.g., which level was positively or negatively associated with higher AMS score.

Line 189: I recommend changing ‘female’ to ‘male’ and reversing the AOR, or changing the description of the association. What is written here implies that being female is associated with ‘good’ AMS attitudes, when in fact the AOR for female was 0.66, so it was ‘protective’ against a ‘good’ AMS attitude (which can be confusing to readers).

Lines 212-213: Similar comment to my prior comment – the authors note that the Western region was a predictor of AMS practices, but the AOR is 0.52, indicating a negative association. Please clarify this in the text to indicate that it was significantly associated with a lower AMS practice score.

Discussion:

Lines 225-231: Are there any other differences between your study and the Ethiopian one that could explain this difference, e.g., years in which the study was carried out, study population, antibiotic availability, etc? Also, this Ethiopian study should probably be mentioned in the introduction as existing data.

Lines 232-233: Please see my note above about indicating the directionality of association. I think it would be clearer to note that male sex was associated with a higher AMS attitude score.

Lines 234-241: The way this association is described is troubling. Firstly, associating sex with good or bad attitudes can be a minefield – see my note above about changing this to high or low scores. Secondly, participant sex is likely confounded by professional cadre, making it difficult to draw inference about this issue. However, the association of participant sex with AMS attitude score was significant while adjusting for level of education, which means that sex is associated with AMS attitude regardless of the level of education. Please examine this interpretation again closely and re-interpret it more carefully.

Additional discussion should be added about why Pharmacists and Pharmacy Techs have high AMS attitude scores but low AMS practice scores – it surprised me to see that their practices were much lower than average and much lower than the other professional cadres.

Additional limitations should be added concerning the lack of data gathered from trainees, and uncertainty around the cause and effect of AMS attitudes and practices.

Conclusion:

The ultimate conclusion – that more needs to be done to support AMS in Uganda – is a good one. However, I don’t think it is justifiable to frame this conclusion in terms of cause and effect – that ‘good’ AMS attitudes are not causing ‘good’ AMS practices – as the authors know, this is association and not causation, and furthermore, there are many systemic barriers to implementing AMS practices regardless of the attitude. This should be acknowledged and the causal language adjusted or removed.

Reviewer #2: This study uses data from interviewer-based surveys administered (from October 2019 to February, 2020) to 582 healthcare providers in 32 (?) facilities to explore attitudes and practices concerning antimicrobial stewardship (AMS) programs in Uganda. The following points will help strengthen the manuscript.

1. Lines 33-50: Could authors add more information to motivate their study of AMS attitudes and practices of healthcare providers? Is this the most important aspect of reducing antibacterial resistance in Uganda? Is the paucity of information on this specific aspect? How was the National Action Plan on Antimicrobial Resistance drafted; bottom-up or top-to-bottom? When drafted? Is it implemented? If yes, how and when (thanks, some information is provided in the Discussion section (lines 227-229) but it needs to be clear from the start) …

2. Lines 53-57: Could authors provide more information about the characteristics of these different hospitals, such as their bed size, teaching status, location characteristics (urban or rural…), their regional distribution etc? Are there any for-profit hospitals in Uganda?

3. Lines 53-57: Could authors explain their reasons for their focus on hospitals only? Are the AMS programs focus only on hospitals?

4. Lines 53-57: Please provide information about the characteristics of regions in Uganda.

5. Lines 69-71: Thanks for the information. This publication requires that at least 3 “sectors” to be identified as a 1st step. So, as before, which are the important “sectors” for the subject matter of this study, only hospitals?

6. Lines 71-74: Thanks for the information. However, it is unclear to the readership how the increase from 4 to 8 regional facilities should be interpreted? The recommendation of “4” is per sector in the publication. Again, are there other “sectors” that are of importance for the subject matter?

7. Lines 74-79: Could authors provide reasons for targeting 24 healthcare providers in each of the hospitals? Are all of these hospitals the same size?

8. Line 82: Again, what are the characteristics of these 4 regions?

9. Lines 82-86: Please be clear if this means that none of the 3 “national referral hospitals” (lines 54-55) were selected into the sample. If so, please provide the rationale.

10. Lines 82-83: Again, how are the 16 regional referral hospitals (line 56) distributed in these 4 regions?

11. Lines 83-85: How are the “50 general (district) hospitals” (line 56) distributed in these 4 regions?

12. Lines 83-85: Please be clear how 3 “general hospitals” in each to the 4 regions add up to “21 general hospitals” (lines 72-73 or line 300) and, therefore, to “32 health facilities” overall (lines 73-74 or Abstract, line 15) for the study?

13. Lines 85-86: Could authors provide information about the regional distribution and characteristics of the 4 “private-not-for-profit health facilities”?

14. Lines 87-88: Please reconcile “departmental heads” here with 4 “heads of departments” above (lines 77-78).

15. Lines 88-90: Again, do all the facilities have the same number of healthcare providers?

16. Lines 130-135: Please be clear if this means that there were no missing data items. If there were missing data items, what procedures were followed?

17. Lines 159-160: Again, please address “768” potential respondents in light of the number of hospitals included in the study sample. As above, what was the number of hospitals in the sample; “32” or 23?

18. Lines 159-218: The reviewer finds it impossible to interpret the results without clarifications for the points above.

19. Lines 227-229: Again, could authors make clear from the very beginning what was the rationale for their study, in general, and the survey, especially the “AMS practice” part, if “the Ministry of Health in Uganda has not implemented any formal interventions on AMS among healthcare providers”?

20. Please avoid typos and ensure completeness and transparency in the manuscript: a) please ensure that all of the acronyms (such as, AMS, in line 25) are spelled out the first time they are mentioned (it is in the Abstract but also needs to be spelled out in the body of the manuscript), b) please ensure that the references are complete (such as #18, for example, please be clear about “WHO” and the location of the publisher…), c) “would; reduce”(?) in line 173 or “is has been”(?) in line 257, d) please ensure that tables are self-explanatory (see, for example, Table 2, what are “AMS” and “60 points” & what does the second note refer to, or Table 4, why “HCP”…)

Reviewer #3: Thank you very much for the opportunity to review this research.

The study investigated healthcare providers' attitude and practices on AMS and associated factors in regional referral hospitals, general hospitals, and private-not-for-profit health facilities in Uganda.

The manuscript is technically sound, and data support the conclusions.

It requires minor corrections. Here are some minor suggestions for improvement.

INTRODUCTION

Please do not start the sentence with an acronym. Also, specify what AMS stands for.

I would suggest organizing the introduction better. I think that it is a little confusing, too many acronyms, and the concepts are mixed. It would be appropriate to outline the context the research refer to and then mention what happened in other countries. Subsequently, I would analyze the problem spread of the antibacterial resistance and the importance of AMS programmes in Uganda.

METHODS

Why did the authors exclude who had worked for only one year in the health facility? What was the rationale?

Authors used ordinal logistic regression to model the data (outcomes were divided into 3 categories). Was the proportional odds assumption checked?

How were variables selected into the models? How was model fit assessed?

DISCUSSION

In the discussion section it seems redundant to define what was done in this study and repeat the results. It is correct instead the comparison with other findings, as the authors did, even though this section should be expanded to present a broader context in which this research is relevant (there are only a few citations to Ethiopia). For instance, results should be interpreted within the perspective of antimicrobial resistance and healthcare-associated infections, considering both low-middle income countries and around the world. Relevant articles worth including may be PMIDs 34223045, 32062724, 33961678, 33882843, 29590400, and 34213520.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

6 Sep 2021

We thank you for offering us an opportunity to share our results with the entire research community.

Submitted filename: Response to reviewers .doc

Click here for additional data file.

5 Oct 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Elena Ambrosino

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: No

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This revised version is significantly improved from prior. I recommend some additional minor changes prior to publication, which I have noted below.

Title and Abstract: Unchanged from prior, no changes recommended.

Introduction: The authors have largely re-written the introduction, and it is well-written. No further changes recommended.

Methods and Results:

Largely re-written. There is excessive detail on the number of hospitals in each region which is somewhat distracting to the reader. The authors still have not described how recruitment/invitation to participate was carried out at each facility to reach the enrollment targets, and whether any compensation/incentive was provided.

In my prior comments, I suggested categorizing the AMS attitude scores as ‘high’, ‘average’, and ‘low’. Re-reading the manuscript now, I think ‘average’ is not the right label, as this has statistical meaning as well. Consider using ‘moderate’ or ‘fair’ to label this category instead of ‘average’.

Results of statistical models have been re-written and are more appropriate. In Line 348, I recommend changing the start of the sentence to ‘In multivariable models adjusting for possible confounders, AMS practice scores …’ This is to clarify that you are referring to your model results. Also include the AOR and its associated confidence interval in this sentence, not just the P-value for the association.

Discussion and Conclusions:

Largely re-written. In line 364 I recommend avoiding the phrase ‘will never’, because this is unlikely to be the case. Consider rephrasing as ‘are unlikely to’. In addition, there is only one limitation addressed – about potential bias in the interview. I think additional limitations should be added concerning the lack of data gathered from trainees (though part-time employees, they do prescribe a significant proportion of antibiotics and have their own AMS attitudes and practices, which were not measured here), uncertainty around the cause and effect of AMS

attitudes and practices, and potential lack of generalizability of your results to non-Ugandan or non-East African settings.

Reviewer #2: Thanks for an improved manuscript. The following needs to be taken into consideration.

1. “Old Lines 33-50: Could authors add more information to motivate their study of AMS attitudes and practices of healthcare providers? Is this the most important aspect of reducing antibacterial resistance in Uganda? Is the paucity of information on this specific aspect? How was the National Action Plan on Antimicrobial Resistance drafted; bottom-up or top-to-bottom? When drafted? Is it implemented? If yes, how and when (thanks, some information is provided in the Discussion section (lines 227-229) but it needs to be clear from the start) …”

Thanks for revisions; the Introduction section motivates the study well. However, it is now more than 3 pages long. Please consider shortening it without losing any of its content.

2. “Old Lines 53-57: Could authors provide more information about the characteristics of these different hospitals, such as their bed size, teaching status, location characteristics (urban or rural…), their regional distribution etc? Are there any for-profit hospitals in Uganda?”

Thanks for providing more information. This section needs to be shortened and, please, just provide information that is relevant to motivate the sampling strategy. Specific questions: (a) Thanks for information about 2 (or is it 3?) “national referral” hospitals (new lines 102-105) that were excluded. As a result, (and names are not required but) what % of the public health hospitals were excluded? (b) “two million” (line 107)? (c) Which region is this; (again, we do not need names and the detail but) why “five regional referral hospitals” followed with 4 names (lines 108-109)? (d) What does “approximately 50 general hospitals” (line 116) mean or are there 41 (lines 115-124) general hospitals? (e) Please be clear, do these hospitals have about 10 beds each (lines 123-124) or 100 beds? (f) If there are “44” private not-for-profit (PNFP) hospitals and if “3 out of 4” that are considered as “regional referral centres” were included in this study, what % of PNFPs was excluded from this study?

3. “Old Lines 53-57: Could authors explain their reasons for their focus on hospitals only? Are the AMS programmes focus at hospitals?”

Thanks for explanations. In the healthcare system, are hospitals the only healthcare institution type where antibacterial misuse is observed?

5. “Old Lines 69-71: Thanks for the information. This publication requires that at least 3 “sectors” to be identified as a 1st step. So, as before, which are the important “sectors” for the subject matter of this study, only hospitals?”

This section in its new format now indicates that information about total number of observational units were available to authors. Was this information also available at the hospital level? If yes, why the sampling was not conducted at the cluster level? Thanks for indicating later that the sample consists mostly of same size “general hospitals” but what about the sizes of other hospitals in the sample (for example is the 1000 bed “regional referral” hospital in the sample)?

6. “Old Lines 71-74: Thanks for the information. However, it is unclear to the readership how the increase from 4 to 8 regional facilities should be interpreted? The recommendation of “4” is per sector in the publication. Again, are there other “sectors” that are of importance for the subject matter?”

Thanks for revisions. If “4 general hospitals for each” of the 8 “regional referral facility” was selected, how did the authors end up with “21 general hospitals” and with the regional distribution of the sample provided (lines 162-165)? Also, see above for what “approximately 50 general hospitals” (line 116) mean.

9. “Old Lines 82-86: Please be clear if this means that none of the 3 “national referral hospitals” (lines 54-55) were selected into the sample. If so, please provide the rationale.”

Thanks for information about 2 “national referral” hospitals (new lines 102-105) that were excluded. Is there a 3rd one?

12. “Old Lines 83-85: Please be clear how 3 “general hospitals” in each to the 4 regions add up to “21 general hospitals” (lines 72-73 or line 300) and, therefore, to “32 health facilities” overall (lines 73-74 or Abstract, line 15) for the study?”

Please see above; “21 general hospitals” is still not clear to the this reviewer.

18. “Old Lines 159-218: The reviewer finds it impossible to interpret the results without clarifications for the points above.”

This reviewer still finds it impossible. Did authors analyze their data as if it was a random sample?

20. Old “Please avoid typos and ensure completeness and transparency in the manuscript: a) please ensure that all of the acronyms (such as, AMS, in line 25), are spelled out the first time they are mentioned (it is in the Abstract but also needs to be spelled out in the body of the manuscript), b) please ensure that the references are complete (such as #18, for example, please be clear about “WHO” and the location of the publisher…), c) “would; reduce”(?) in line 173 or “is has been”(?) in line 257, d) please ensure that tables are self-explanatory (see, for example, Table 2, what are “AMS”, “60 points” and what does the second note refer to, or Table 4, why “HCP”…)”

Some of the new points: a) new line 27, do you mean “arises”, b) new line 33, the sentence that starts with “Antimicrobial stewardship (AMS)…” needs to be a new paragraph, c) new line 61, “where” seems to be missing, d) “LMIC”? new line 76….

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

15 Nov 2021

RESPONSE TO REVIEWER'S COMMENTS ON MANUSCRIPT, PONE-D-21-10939R1

Dear Editor and reviewers:

We greatly appreciate the thorough and thoughtful comments provided on our submitted manuscript. We have revised our manuscript according to the reviewers' comments, questions, and suggestions. Please see below our detailed responses to the reviewers' comments. All our responses are in bold. Please let us know if you still have any questions or concerns about the manuscript, and we will be happy to address them promptly.

Yours,

The authors of PONE-D-21-10939R1

Editor's comment:

Thank you to the authors for their work on revising the manuscript so far. Could the authors now focus on the point below, explaining what choices they made when administering the surveys and clarifying the rationale behind such choices?

Response to Editor's comment #1:

We choose different types of health facilities for this study because of a gap in the literature on attitudes and practices of healthcare providers between urban regional referral hospitals, private not for profit (PNFP) hospitals and community general hospitals. Previous studies in Nigeria and Ethiopia had concentrated on studying only tertiary or specialised hospitals, limiting their generalizability. The study hospitals were selected due to the presence of structures and systems for strengthening the monitoring and improvement of medicine use. Most government and development partners capacity building has focused on establishing infection prevention committees, medicines and therapeutic committees in these hospitals.

We chose a large sample size of healthcare providers to increase the study's representativeness and determine the predictors of AMS attitude and practices in health facilities in Uganda.

We chose the administration office and heads of department to recruit the healthcare providers into our study. We used emails, letters, and phone calls to recruit healthcare providers to increase the participation of healthcare providers.

Every participant consented first before being interviewed.

We used an interview-administered questionnaire. We used research assistants from the same facility to administer the questionnaire to minimise bias in the responses. During piloting, we realised respondents gave different responses when the principal investigator or research assistant interviewed, so we used a research assistant other than the investigator himself to minimise the measurement bias. We used research assistants from different health backgrounds within the health facility to avoid the interviewer social desirability effect.

Editor's comment:

The manuscripts provides a good amount of information about, for example, the healthcare worker population, hospitals, and regions. It does not appear as though this info has been used to devise an appropriate cluster sampling, which would have in turn guided the analysis of data.

Response to Editor's comment #2:

However, much as we couldn't conduct cluster sampling, the study catered for clustering at the healthcare provider level at the health facility

Editor's comment:

Although the authors calculated an overall sample size, this was not used to administer the surveys randomly around the country, as the administration ultimately does not appear random, but rather distributed to certain (types of) hospitals in some regions.

Response to Editor's comment #3:

Yes, we agree with the Editor's comments. We could not use random sampling to all types of hospitals since we had to receive administrative clearance to know the sampling frame of participating hospitals. We randomly selected eight out of 16 regional referrals, where two regional referrals were picked per region to increase regional representativeness. All eight out 16 had granted us administrative clearance. However, of all the 32 out of 47 general hospitals selected after random sampling, only 21 granted us administrative clearance to conduct the study. Eleven general hospitals didn't participate in the selected sample.

We could not randomly select pharmacists, pharmacy technicians, and laboratory technicians in all hospitals that participated since each hospital had only one employee of these types of professional cadre. However, nurses, medical officers, clinical officers and medical specialists were selected randomly.

Reviewer #1

This revised version is significantly improved from prior. I recommend some additional minor changes prior to publication, which I have noted below.

Response:

Thanks for the valuable comments, and we appreciate your time.

Methods and Results:

Largely re-written. There is excessive detail on the number of hospitals in each region which is somewhat distracting to the reader.

Response:

Lines 99-107 on page 6; We have improved the clarity of writing in this section, deleted names of hospitals and limited ourselves to the number of targeted hospitals and the reasons for their selection.

The authors still have not described how recruitment/invitation to participate was carried out at each facility to reach the enrollment targets and whether compensation/incentive was provided.

Response:

Lines 197-205, page 11. We apologise for that, and we have added more information on how the research assistant conducted recruitment. We have also added more information on compensation for time and transport since some respondents were on leave and were requested to come to the facility

In my prior comments, I suggested categorising the AMS attitude scores as 'high', 'average', and 'low'. Re-reading the manuscript now, I think 'average' is not the right label, as this has statistical meaning as well. Consider using 'moderate' or 'fair' to label this category instead of 'average'.

Response:

The authors want to thank the review and have agreed to use "fair" instead of "average", as reflected in this manuscript's sections.

Comment Reviewer #1:

Results of statistical models have been re-written and are more appropriate. In Line 348, I recommend changing the start of the sentence to 'In multivariable models adjusting for possible confounders, AMS practice scores …' This is to clarify that you are referring to your model results. Also include the AOR and its associated confidence interval in this sentence, not just the P-value for the association.

Response:

Lines 321-323, page 23-24. We have corrected it as suggested.

Comment Reviewer #1:

Discussion and Conclusions:

Largely re-written. In line 364 I recommend avoiding the phrase 'will never', because this is unlikely to be the case.

Response:

Lines 333-334, pages 25, Thanks for the correction. We have improved the clarity of the statement

Consider rephrasing as 'are unlikely to'. In addition, there is only one limitation addressed – about potential bias in the interview. I think additional limitations should be added concerning the lack of data gathered from trainees (though part-time employees, they do prescribe a significant proportion of antibiotics and have their own AMS attitudes and practices, which were not measured here), uncertainty around the cause and effect of AMS attitudes and practices, and potential lack of generalizability of your results to non-Ugandan or non-East African settings.

Response:

Lines 417-419, page 29. We have included the limitation of part and intern healthcare providers.

Lines 422-423, Pages 29. We have added a limitation on the uncertainty surrounding the cause and effect of AMS attitudes and practices.

Lines 423-424, Pages 29: We have added an aspect on our results not being generalisable to other non-East African settings or non-Ugandan setting

Reviewer #2

Reviewer #2: Thanks for an improved manuscript. The following needs to be taken into consideration.

Response:

Thanks for the valuable comments; we appreciate

Comment Reviewer #2:

Thanks for revisions; the Introduction section motivates the study well. However, it is now more than three pages long. Please consider shortening it without losing any of its content.

Response:

Lines 26-92, We have shortened the introduction section and its now two pages.

Comment Reviewer #2:

Thanks for information about 2 (or is it 3?) "national referral" hospitals (new lines 102-105) that were excluded. a)As a result, (and names are not required but) what % of the public health hospitals were excluded?

Response:

Lines 86-107, pages 6. We have clarified the health system characteristics. In our clarification, we report that the health system comprises 6937 health facilities, and of these, 3133 (45.6%) are public health facilities. The composition of the latter includes two national referral hospitals, three referral hospitals (the Uganda Cancer Institute, the Uganda Heart Institute, and the Women's Hospital), 13 regional referral hospitals, 47 general hospitals, 166 level IV health centres, 962 level III health centres, and 1321 level II health centres. We had only 63 out of 3133 (2%) health facilities participating. We excluded 3070 health facilities. The percentage of health facilities excluded is 98% (3070/3133).

(a) "two million" (line 107)?

Response:

Lines 86-107, pages 6. Thanks for this highlight. According to the health facilities of Uganda Master list (2018) page 7, Each health facility has a designated population size that it is meant to serve; a national referral hospital (10,000,000 persons), each regional referral hospital (2,000,000 persons), general hospitals (500,000) and, Health centre IV (100,000), Health centre III (20000), health centre II (5000) and health centre I (not defined).

In our previous description of the study setting, we had included this characteristic showing that a regional referral serves over two million persons; however, we have removed it in the revised manuscript for clarity.

(c) Which region is this; (again, we do not need names and the detail but) why "five regional referral hospitals" followed with 4 names (lines 108-109)?

Response:

Lines 99-107, pages 6. We thank you for this observation and the need for clarification. In the central region, we have five regional referrals hospitals that we included in our study.

Comment Reviewer #2:

(d) What does "approximately 50 general hospitals" (line 116) mean or are there 41 (lines 115-124) general hospitals? (e) Please be clear, do these hospitals have about 10 beds each (lines 123-124) or 100 beds? (f) If there are "44" private not-for-profit (PNFP) hospitals and if "3 out of 4" that are considered as "regional referral centres" were included in this study, what % of PNFPs was excluded from this study?

Response:

Lines 95-107, page 6. In the PNFP health systems, there are 1009 health facilities with the highest referral level in the four tertiary hospitals, followed by 40 general hospitals and 955 health centres. We excluded all the 40 general PNFP hospitals and 955 health centres making a percentage of 98.6%.

Comment Reviewer #2:

3. "Old Lines 53-57: Could authors explain their reasons for their focus on hospitals only? Are the AMS programmes focus at hospitals?"

Thanks for explanations. In the healthcare system, are hospitals the only healthcare institution type where antibacterial misuse is observed?

Response:

Lines 102-108, page 6. Thanks for the comment. We agree with you that overuse and misuse of antibacterials are not limited to regional referrals, general hospitals and PNFPs. Recent health initiatives have focused on training most public health facilities, tertiary PNFP healthcare providers, and private healthcare providers in communities on antimicrobial resistance and antimicrobial stewardship. Indeed most healthcare providers have been trained on the implementation of antimicrobial stewardship interventions. There has been the ongoing implementation of the National action plan, and the focus of the government has been on strengthening existing or initiating medicines and therapeutic committees. The committees implement antimicrobial stewardship, pharmacovigilance, and supply chain management in regional referrals, general hospitals, and PNFPs. Our study focused on assessing the attitudes and practices of healthcare providers on the implementation of these programs. However, further initiatives are needed to understand the attitudes and practices of healthcare workers in the lower health facilities since there are many challenges like infrastructure, human resources and follow of information.

Comment Reviewer #2:

This section in its new format now indicates that information about total number of observational units were available to authors. Was this information also available at the hospital level?

Response:

Lines 130-141, page 8. Thanks for this comment. We could not know how many hospitals we shall collect from data not until the hospitals granted ethical and administrative clearance. We did not have information at the hospital level because the ethical guidelines in Uganda required the researcher to be issued ethical clearance from the parent institution and national ethical clearance from the national institution. Finally, administrative clearance to all respective hospitals we targeted before being availed of any information at the hospital level. No respondent can participate in the study without any administrative clearance. After selecting eight out of 16 regional referrals, clearance was granted by all the selected eight regional referrals hospitals. After selecting 32 out of the 47 general hospitals, administrative clearance was granted by only 21 hospitals.

If yes, why the sampling was not conducted at the cluster level?

Response:

Lines 130-141, page 8. The sampling procedure did not take a cluster sampling approach because of a shortfall in the participation of hospitals in the northern region of the country (seven general hospitals did not grant us administrative clearance). The cluster of North would have had one general hospital compared to six general hospitals in the Central, seven general hospitals in Eastern and seven in the western. There was an imbalance in terms of general hospitals. The North would not form a cluster because of the non-participation of six hospitals. Therefore, there would be a shortfall in the study population (healthcare providers) representation northern region. Furthermore, this would mean there would be one cluster with no similar characteristics in terms of healthcare providers or participating types of hospitals compared to other regions.

However, since there was clustering among healthcare providers in terms of cadres at the lowest level, the study adjusted for clustering in the design effect of sample size determination also adjusted for clustering during analysis using robust standard errors.

Comment Reviewer #2:

Thanks for indicating later that the sample consists mostly of same size "general hospitals" but what about the sizes of other hospitals in the sample (for example is the 1000 bed "regional referral" hospital in the sample)?

Response:

Lines 143-149, page 8. Thanks for this comment. The study ensured proportionate representation among participating hospitals by using a proportionate number to size to select all other professional cadres except pharmacists, pharmacy technicians and laboratory technicians because each hospital had one cadre. The latter were purposively selected in each participating hospital.

Comment Reviewer #2: If "4 general hospitals for each" of the 8 "regional referral facility" was selected, how did the authors end up with "21 general hospitals" and with the regional distribution of the sample provided (lines 162-165)? Also, see above for what "approximately 50 general hospitals" (line 116) mean.

If "4 general hospitals for each" of the 8 "regional referral facility" was selected, how did the authors end up with "21 general hospitals"

Response:

Lines 133-141, page 8. Thanks for this question. Before data collection, we selected randomly eight regional referrals hospitals. We selected 32 general hospitals out of the 47 included general hospitals. For each of the selected regional referrals, we had to select four general hospitals. Therefore for all the four regions (North, West, East and Central), we selected 32 general hospitals out of the 47. How did we end up with 21 general hospitals? After selecting the eight regional referrals, 32 general hospitals, and three tertiary PNFPs, we applied for administrative clearance in each selected health facility. All the eight regional referrals and three tertiary PNFPs granted us administrative clearance. However, for general hospitals, only 21 out of 32 granted us administrative clearance?. The ethical obligation by the Uganda National Council of Science and Technology and the parent institution (Makerere University Higher degrees ethics and research committees and the hospital research and ethics committee) can only be done once granted ethical clearance by all the three institutions. That is how 21 general hospitals appear in the stated methods.

Also, see above for what "approximately 50 general hospitals" (line 116) mean.

Lines 91-92, page 6. Thanks for this question, and we apologies for not being precise on using "approximately 50" general hospitals. According to the national census for health facilities carried out in 2014 and the national health facility master list, there are 47 general hospitals in Uganda.

Comment Reviewer #2:

Thanks for information about 2 "national referral" hospitals (new lines 102-105) that were excluded. Is there a 3rd one?

Response:

Lines 91-92, page 6. By 2014 when a national census for health facilities was done, there were three national referrals, Mulago super specialised hospitals, Butabika national referral for mental health, and Buluba national referral for leprosy. In the latest health facility master list 2018, the Ministry of Health no longer recognises Buluba as a national referral, and it implies we are left with two national referral hospitals.

Comment Reviewer #2:

Please see above; "21 general hospitals" is still not clear to the this reviewer.

Response:

Lines 144-156, page 8. Thanks for this question. Before data collection, we selected randomly eight regional referrals. Then we selected 32 of the 47 general hospitals, where for each of the selected regional referrals, we had to select four general hospitals. We selected 32 general hospitals out of the 47. How did we end up with 21 general hospitals? After selecting the 32 general hospitals, we applied for administrative clearance in each of the selected general hospitals. Only 21 general hospitals granted us the clearance to conduct the study with them. These included one in the North, six in the central, seven in the East and seven in the West. The ethical obligation to the Uganda National Council of Science and Technology and the Makerere University Higher degrees ethics and research committee, and the hospital research and ethics committee was that the researcher could only conduct data collection after all three institutions had received ethical clearance. That is how the 21 general hospitals appear stated in our methods.

18. "Old Lines 159-218: The reviewer finds it impossible to interpret the results without clarifications for the points above."

Response :

Lines 132-141, page 8. After clarifying the information on the number of hospitals that participated and how they were selected we believe the reviewer will agree with us on this revision that results can be easily interpreted

Comment Reviewer #2:

This reviewer still finds it impossible. Did authors analyse their data as if it was a random sample?

Response:

The answer is a yes or no. We had chosen facilities randomly, but the healthcare providers like pharmacists, laboratory technicians and pharmacy technicians were chosen purposively since we only found one person per health facility. In contrast, nurses, medical officers, medical specialists, and clinical officers all were chosen by simple random sampling using sampling frames of the hospital facilities. Having a sample with a mixture of persons who were picked purposively and others randomly the answer may lie most on the no side.

Comment Reviewer #2:

20. Old "Please avoid typos and ensure completeness and transparency in the manuscript:

Response

Thanks for this observation, we have corrected all typos in the manuscript

a) please ensure that all of the acronyms (such as, AMS, in line 25), are spelled out the first time they are mentioned (it is in the Abstract but also needs to be spelled out in the body of the manuscript),

Response

Lines 2, page 1 and Lines 36, page 2. Thanks for the correction, We have spelt out the word AMS twice in the abstract and once in the main section

b) please ensure that the references are complete (such as #18, for example, please be clear about "WHO" and the location of the publisher…),

Response

Lines 498-601: Thanks for the correction, and we have ensured the references are all complete.

c) "would; reduce" (?) in line 173 or "is has been" (?) in line 257, d) please ensure that tables are self-explanatory (see, for example, Table 2, what are "AMS", "60 points" and what does the second note refer to, or Table 4, why "HCP"…)"

Response

Lines 271 pages 16. Thanks for the correction, "would reduce" corrected to "minimises."

Lines 390 pages 27. "is has been" corrected to "has been".

Table 2: We reported the mean score in 12 questions on antimicrobial stewardship attitudes. This has been improved.

Table 4: Sorry for the HCP. It has been corrected.

Some of the new points:

a) new line 27, do you mean "arises",

Response

Lines 28-29: we corrected the statement.

b) new line 33, the sentence that starts with "Antimicrobial stewardship (AMS)…" needs to be a new paragraph,

Response

Lines 37: We thank you for this suggestion. We have made this correction.

b) new line 61, "where" seems to be missing, d) "LMIC"? new line 76….

Lines 73, pages 2: we have corrected the abbreviation

________________________________________

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose "no", your identity will remain anonymous but your review may still be made public.

No

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Response to reviewers .docx

Click here for additional data file.

23 Nov 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Elena Ambrosino

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

6 Jan 2022

I thank you for the time and comments on this manuscript, they have improved the manuscript.

Submitted filename: response to reviewers.docx

Click here for additional data file.

11 Jan 2022

Antimicrobial stewardship: Attitudes and practices of healthcare providers in selected health facilities in Uganda

PONE-D-21-10939R3

Dear Dr. Kimbowa,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Elena Ambrosino

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

25 Jan 2022

PONE-D-21-10939R3

Antimicrobial stewardship: Attitudes and practices of healthcare providers in selected health facilities in Uganda

Dear Dr. Kimbowa:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Elena Ambrosino

Academic Editor

PLOS ONE