Infection prevention and control practices among dental healthcare professionals in Somaliland: a multicity cross-sectional study

Abstract

Background

Dental healthcare personnel are exposed to blood, saliva, sharp instruments, and procedure-generated aerosols. Evidence on infection prevention and control (IPC) practices in Somaliland dental services is limited. This study assessed IPC practices and examined factors associated with the IPC practice score among dental healthcare professionals.

Methods

A facility-based cross-sectional survey included 180 dental healthcare professionals from accessible public and private dental facilities in four Somaliland study areas. A structured self-administered questionnaire measured professional characteristics, IPC systems, vaccination, knowledge, attitudes, reported practices, occupational exposure, and barriers. Five knowledge items produced a score of 0–5. Seven frequency-based practice items were scored from 0 to 4 and summed to a score of 0–28; scores of at least 75% were classified as good practice. Univariable and multivariable linear regression with robust standard errors estimated associations with the continuous practice score. Fractional-logit and workload-restriction analyses assessed sensitivity.

Results

Participants had a mean age of 27.2 years (SD 3.5); 64.4% were male and 74.4% were general dentists. The mean knowledge score was 3.20/5 (SD 1.00), and 33.9% had good knowledge. The mean practice score was 25.73/28 (SD 2.34), and 174/180 (96.7%) reached the prespecified 75% threshold. Always performing hand hygiene before and after each patient was reported by 81.1%; always using and changing gloves between patients was reported by 96.7%. Only 21.7% selected an N95 respirator for aerosol procedures, and 22.2% reported biological-indicator sterilization monitoring. In adjusted analysis, no recent IPC training (adjusted beta -1.44, 95% CI −2.47 to −0.41), PPE supplies available only sometimes (−1.38, 95% CI −2.42 to −0.34) or rarely (−4.99, 95% CI −7.02 to −2.96), and treating 10–19 rather than fewer than 10 patients daily (−0.89, 95% CI −1.67 to −0.11) were associated with lower practice scores. Sensitivity analyses produced materially similar patterns.

Conclusions

Reported compliance with several routine practices was high, but important gaps remained in respiratory protection, sterilization monitoring, occupational vaccination, post-exposure systems, and reliable PPE supply. Because the total eligible population and response rate were unknown, and practices were self-reported, the magnitude and generalizability of these estimates remain uncertain. Facility-level IPC programs should prioritize dependable supplies, recurrent competency-based training, sterilization quality assurance, vaccination, and post-exposure management.

Background

Healthcare-associated infections harm patients and health workers, increase health-service costs, and contribute to antimicrobial resistance. Their burden is amplified where infection prevention and control (IPC), water, sanitation, hygiene, staffing, and surveillance systems are constrained [1, 2]. Dental care requires particular attention because clinical procedures combine close face-to-face contact with blood, saliva, contaminated instruments, sharps, droplets, and aerosols. These exposures create opportunities for direct, indirect, droplet, airborne, and percutaneous transmission [3–5].

Standard Precautions are the minimum practices required for every patient encounter. In dental settings, they include hand hygiene, appropriate personal protective equipment (PPE), safe handling and sterilization of reusable instruments, environmental cleaning, sharps safety, vaccination, and prompt post-exposure management [3, 4]. CDC guidance also recommends written facility-specific procedures, a trained IPC coordinator, continuing education, adequate supplies, and routine evaluation of the IPC program [3, 4]. These elements place responsibility on both individual practitioners and the systems in which they work.

The consequences of incomplete dental IPC extend beyond individual patient encounters. Failures in instrument reprocessing, respiratory protection, sharps management, immunization, or post-exposure care can expose patients, clinicians, support staff, and their households to avoidable infection risks. In resource-constrained services, the effectiveness of recommended precautions may depend as much on procurement, facility organization, monitoring, and workforce support as on individual knowledge [1, 2, 6]. Evaluating routine behavior without examining these enabling systems can therefore produce an incomplete account of IPC performance.

Evidence from low- and middle-income settings shows that knowledge alone does not ensure implementation. Compliance is influenced by training, supply availability, workload, safety culture, supervision, and facility infrastructure [6–11]. Dental surveys have similarly documented uneven performance across hand hygiene, PPE, sterilization, environmental control, vaccination, and occupational-exposure management [7]. Much of the available literature, however, comes from health systems with established professional registers, facility inventories, or formal IPC programs. Its direct applicability to Somaliland—where dental services are delivered through a mixture of private clinics, public hospital dental units, teaching clinics, and non-governmental facilities cannot be assumed.

Published empirical evidence describing dental IPC in Somaliland is scarce. This evidence gap limits the ability of facility managers, training institutions, professional bodies, and health authorities to distinguish highly adopted routine precautions from technically demanding or system-dependent weaknesses. It also leaves uncertainty about whether reported practice varies with recent IPC training, reliable PPE availability, patient workload, written procedures, designated IPC responsibility, or access to post-exposure protocols. Locally generated evidence is therefore needed to define practical priorities, while avoiding claims of national representativeness when the underlying workforce denominator is unknown.

This study assessed reported IPC knowledge, facility systems, material resources, occupational exposure, perceived barriers, and practice among dental healthcare professionals in selected Somaliland settings and examined factors associated with a seven-item IPC practice score. Its principal contribution is an integrated item-level and adjusted analysis that separates commonly reported routine precautions from gaps requiring facility- or health-system action. The findings are intended to inform dental IPC training, procurement, sterilization quality assurance, vaccination, injury reporting, and post-exposure planning, while explicitly accounting for the incomplete sampling frame, self-reported measurement, and absence of participant-level city identifiers.

Methods

Study design and setting

A facility-based cross-sectional survey was conducted in accessible dental facilities in Hargeisa, Burao, Erigavo, and the Awdal study area in Somaliland. Public hospital dental units, private clinics, teaching clinics, and non-governmental organization clinics were eligible. The original field records did not include a participant-level city variable in the analytical dataset; consequently, the cities describe the intended fieldwork scope, but city-stratified estimates could not be produced. Data collection occurred over a 3-month period; exact dates were not available in the submitted study records and are therefore not stated.

Population and eligibility

The study population comprised dental healthcare professionals directly involved in clinical care. Eligible personnel included general dentists, dental specialists, dental therapists, dental assistants, dental hygienists, and other clinical dental personnel. Administrative and support staff without direct clinical duties, personnel employed for less than 3 months, and staff absent during data collection were excluded. Written informed consent was obtained before participation.

Sampling and sample size

A census approach was attempted among eligible personnel working in accessible facilities during the study period. All accessible eligible personnel were approached, and 180 completed questionnaires were analyzed. A comprehensive registry of all dental facilities and eligible dental healthcare professionals in the four study areas was unavailable. The total number eligible and the number approached could therefore not be verified, and a response rate could not be calculated. Accordingly, the sample should be interpreted as a census of accessible participants rather than a documented census of the underlying workforce.

Instrument and data collection

Data were collected using a structured, self-administered questionnaire informed by CDC dental IPC recommendations [3, 4]. It covered sociodemographic and professional characteristics; IPC training and facility systems; vaccination; hand hygiene; PPE; sterilization and monitoring; environmental cleaning and ventilation; waste management; needle-stick exposure and reporting; five knowledge items; attitudes; and perceived barriers. Knowledge items had True, False, and Not sure response options. Four attitude statements used a five-point agreement scale; the separately coded statement on whether guidelines slow work used Yes, No, and Not sure and was not included in the four-item attitude score. Practice items used frequency responses. The instrument was pretested outside the study facilities, revised for clarity, administered by trained data collectors, and checked for completeness at collection.

Variables and scoring

The primary outcome was the seven-item IPC practice score. Items measured hand hygiene before and after each patient, routine glove use, changing gloves between patients, packaging instruments before sterilization, disinfection of chairs and surfaces between patients, high-volume evacuator use during aerosol procedures, and disposal of sharps in puncture-proof containers. For five-category items, Always, Often, Sometimes, Rarely, and Never were scored 4, 3, 2, 1, and 0. The packaging item contained four recorded options and was scored Always = 4, Often = 3, Sometimes = 2, and Never = 0. Complete-item scores ranged from 0 to 28. A score of at least 21 (75% of 28) was classified as good practice for descriptive reporting.

Five knowledge statements were scored using a prespecified scientific answer key: True for complete microorganism elimination by sterilization, False for alcohol hand rub when hands are visibly soiled, True for hepatitis B transmission during dental procedures, False for full airborne protection from surgical masks, and True for biological indicators as the sterilization-monitoring gold standard. Correct responses received one point; incorrect and Not sure responses received zero. The total ranged from 0 to 5 and was categorized as poor (<60%; 0–2), moderate (60–79%; 3), or good (80–100%; 4–5). The limited five-item breadth was considered when interpreting knowledge and internal consistency.

The four-item favorable-attitude score ranged from 0 to 16 after aligning response direction. Counts were also calculated for six PPE types selected for aerosol procedures (0–6) and seven perceived barriers (0–7). Explanatory variables included age, sex, professional category, clinical experience, facility type, patients treated per day, IPC training during the previous two years, written guidelines, designated IPC officer, hand-hygiene and PPE supply availability, knowledge score, and post-exposure protocol availability.

Data management and statistical analysis

The original Excel export was re-imported and checked against the questionnaire, SPSS metadata, and raw Stata records. Raw response fields were retained. One value of 1999 entered in the age field was preserved as age_raw and converted to 27 years using the 2026 collection year. Two identical response profiles with different UUIDs were flagged but retained. One participant had missing barrier items. Two patient-load values (150 and 300 patients/day) were retained in the primary analysis and excluded in a sensitivity analysis.

Analyses were conducted in Stata 17. Frequencies and percentages summarized categorical variables. Means with standard deviations and medians with interquartile ranges summarized continuous variables. Histograms, normal Q-Q plots, and Shapiro-Wilk tests assessed distributions. Cronbach alpha described internal consistency but was interpreted cautiously because each domain contained few and heterogeneous items. Pearson and Spearman correlations and nonparametric group comparisons were exploratory.

Univariable linear regression with robust standard errors estimated crude associations with the 0-28 practice score. A prespecified multivariable linear model included age, sex, professional category, clinical experience, facility type, patient-load category (<10, 10–19, ≥20), IPC training, written guidelines, designated officer, hand-hygiene supplies, PPE supplies, knowledge score, and post-exposure protocol. Results are presented as crude and adjusted beta coefficients with 95% confidence intervals. Variance-inflation factors assessed multicollinearity; residual plots, the Breusch-Pagan/Cook-Weisberg test, Ramsey RESET, and influence statistics assessed model assumptions. Robust standard errors were retained because heteroscedasticity was detected. Fractional-logit modeling of the proportion of the maximum practice score and exclusion of patient-load values above 100 were sensitivity analyses. Binary or ordered logistic regression was not fitted because only six participants were below the 75% threshold. ROC analysis was not applicable because the primary outcome was a continuous practice score, not a diagnostic classifier. Statistical tests were two-sided with p < 0.05, while interpretation emphasized magnitude, precision, plausibility, and sparse categories. Reporting followed STROBE guidance [12].

Ethical considerations

Ethical approval was obtained from the Institutional Review Board of Burao Central University College, College of Health Science, Department of Dentistry (reference BCU/CHS/DD/58/26). Participants received information about study procedures, voluntary participation, confidentiality, and withdrawal and provided written informed consent. Study procedures were conducted in accordance with the Declaration of Helsinki.

Results

Participant characteristics

All 180 completed questionnaire records were retained in the corrected analysis; no record was excluded. The number eligible and approached was unavailable because a complete workforce and facility registry did not exist, so a response rate could not be reconstructed (Fig. 1). Participants had a mean age of 27.2 years (SD 3.5), a median of 27 years (IQR 25–28), and a range of 21–40 years. The median number of patients treated per day was 12 (IQR 8–17.25; range 1–300). Most participants were male (64.4%), general dentists (74.4%), bachelor’s degree holders (79.4%), and employed in private clinics (82.8%) (Table 1). Participant-level city was not available, so regional distributions and comparisons could not be calculated.

Fig. 1. Participant flow for the corrected analysis.

The total eligible and approached denominator could not be reconstructed because no complete workforce and facility registry was available. All 180 completed questionnaire records were retained in the corrected analysis.

Table 1.

Sociodemographic and professional characteristics of participants (N = 180).

Characteristic	Category/statistic	Value
Age, years	Mean (SD)	27.2 (3.5)
Age, years	Median (IQR)	27 (25–28)
Age, years	Range	21–40
Sex	Male	116 (64.4)
	Female	64 (35.6)
Professional category	General dentist	134 (74.4)
	Dental specialist	23 (12.8)
	Dental therapist	2 (1.1)
	Dental assistant	19 (10.6)
	Other	2 (1.1)
Education	Diploma	5 (2.8)
	Bachelor’s degree	143 (79.4)
	Master’s degree	28 (15.6)
	Specialist certification	3 (1.7)
	Other	1 (0.6)
Clinical experience	Less than 1 year	19 (10.6)
	1–5 years	121 (67.2)
	6–10 years	31 (17.2)
	More than 10 years	9 (5.0)
Facility type	Private clinic	149 (82.8)
	Public hospital dental unit	25 (13.9)
	NGO clinic	1 (0.6)
	Teaching clinic	5 (2.8)
Patients treated/day	Median (IQR)	12 (8–17.25)
	Range	1–300

Values are n (%) unless otherwise indicated.

IQR interquartile range, SD standard deviation.

IPC systems, vaccination, and occupational exposure

Formal IPC training during the previous 2 years was reported by 130 participants (72.2%). Written IPC guidelines were reported as available by 138 (76.7%), and 128 (71.1%) reported a designated IPC officer. Hepatitis B vaccination was reported by 131 participants (72.8%), COVID-19 vaccination by 97 (53.9%), and tetanus vaccination by 91 (50.6%). A post-exposure protocol was reported as available by 119 (66.1%). Fifty-eight participants (32.2%) reported a needle-stick injury in the previous 12 months; among them, 44 (75.9%) reported the injury (Table 2).

Table 2.

IPC training, clinic systems, vaccination, and occupational exposure.

Variable	Response	n (%)
Formal IPC training in previous 2 years	Yes	130 (72.2)
	No	50 (27.8)
Written IPC guidelines available	Yes	138 (76.7)
	No	34 (18.9)
	Not sure	8 (4.4)
Designated IPC officer	Yes	128 (71.1)
	No	52 (28.9)
Hepatitis B vaccination	Yes	131 (72.8)
	No	45 (25.0)
	Not sure	4 (2.2)
COVID-19 vaccination	Yes	97 (53.9)
	No	80 (44.4)
	Not sure	3 (1.7)
Tetanus vaccination	Yes	91 (50.6)
	No	78 (43.3)
	Not sure	11 (6.1)
Post-exposure protocol available	Yes	119 (66.1)
	No	41 (22.8)
	Not sure	20 (11.1)
Needle-stick injury in previous 12 months	Yes	58 (32.2)
	No	122 (67.8)
Injury reported among exposed participants (n = 58)	Yes	44 (75.9)
	No	14 (24.1)

Percentages use N = 180 except injury reporting, for which the denominator is the 58 participants reporting a needle-stick injury.

Reported IPC practices

Always performing hand hygiene before and after each patient was reported by 146 participants (81.1%). Routine glove use and changing gloves between patients were each reported as always by 174 (96.7%). Instrument packaging was reported as always by 149 (82.8%), chair or surface disinfection by 147 (81.7%), and puncture-proof sharps disposal by 138 (76.7%). High-volume evacuator use was less consistent: 95 (52.8%) reported always using one and 15 (8.3%) reported never using one (Table 3).

Table 3.

Distribution of responses to the seven IPC practice items (N = 180).

Practice item	Always	Often	Sometimes	Rarely	Never
Hand hygiene before and after each patient	146 (81.1)	13 (7.2)	19 (10.6)	2 (1.1)	0 (0.0)
Routine glove use during procedures	174 (96.7)	5 (2.8)	1 (0.6)	0 (0.0)	0 (0.0)
Change gloves between patients	174 (96.7)	5 (2.8)	1 (0.6)	0 (0.0)	0 (0.0)
Package instruments before sterilization	149 (82.8)	16 (8.9)	14 (7.8)	NA	1 (0.6)
Disinfect chair/surfaces between patients	147 (81.7)	22 (12.2)	8 (4.4)	3 (1.7)	0 (0.0)
Use high-volume evacuator for aerosol procedures	95 (52.8)	33 (18.3)	30 (16.7)	7 (3.9)	15 (8.3)
Dispose sharps in puncture-proof containers	138 (76.7)	23 (12.8)	15 (8.3)	1 (0.6)	3 (1.7)

Values are n (%). Instrument packaging had four recorded options (Always, Often, Sometimes, Never); Rarely was not available for that item. Scoring: Always = 4, Often = 3, Sometimes = 2, Rarely = 1, Never = 0; packaging Never = 0. Composite score: mean 25.73 (SD 2.34), median 26 (IQR 24–28), range 17–28. At least 75%: 174/180 (96.7%). NA, not applicable.

The seven-item practice score was concentrated near the upper limit, with a mean of 25.73 (SD 2.34), median of 26 (IQR 24–28), and range of 17–28. Overall, 174 participants (96.7%) met the prespecified threshold of at least 75%. The histogram and Q-Q plot show the left-skewed, ceiling-constrained distribution (Supplementary Fig. S1A, B).

PPE, sterilization, environmental controls, and waste

For aerosol procedures, 151 participants (83.9%) selected a surgical mask, 127 (70.6%) a gown, 88 (48.9%) a face shield, 83 (46.1%) protective eyewear, 39 (21.7%) an N95 respirator, and 36 (20.0%) a cap. Autoclaving was the most frequently reported sterilization method (62.2%). Sterilization monitoring relied on mechanical indicators for 37.8%, chemical indicators for 32.2%, biological indicators for 22.2%, while 7.8% reported no monitoring. Adequate ventilation was reported by 80.6%, and waste segregation at source by 91.7% (Table 4).

Table 4.

PPE selection, sterilization, environmental controls, and waste management (N = 180).

Domain	Response	n (%)
Aerosol PPE selected	Surgical mask	151 (83.9)
	N95/respirator	39 (21.7)
	Face shield	88 (48.9)
	Protective eyewear	83 (46.1)
	Gown	127 (70.6)
	Cap	36 (20.0)
Sterilization method	Autoclave	112 (62.2)
	Dry heat sterilizer	53 (29.4)
	Chemical sterilizer	7 (3.9)
	Boiling	7 (3.9)
	Other	1 (0.6)
Sterilization monitoring	Biological indicator	40 (22.2)
	Chemical indicator	58 (32.2)
	Mechanical indicator	68 (37.8)
	None	14 (7.8)
Adequate ventilation	Yes	145 (80.6)
	No	24 (13.3)
	Not sure	11 (6.1)
Waste segregated at source	Yes	165 (91.7)
	No	15 (8.3)
Final waste-disposal method	Incineration	50 (27.8)
	Open burning	43 (23.9)
	Municipal waste	55 (30.6)
	Other	32 (17.8)

PPE selections were multiple-response items; percentages therefore do not total 100%. Mean PPE selection count was 2.91 (SD 1.51), median 3 (IQR 1–4), range 1–6.

Knowledge, attitudes, and barriers

The mean knowledge score was 3.20/5 (SD 1.00), with a median of 3 (IQR 3–4) and range of 0–5. Knowledge was poor for 33 participants (18.3%), moderate for 86 (47.8%), and good for 61 (33.9%). The least frequently correct statements concerned the suitability of alcohol hand rub when hands are visibly soiled (31.7% correct) and full airborne protection from surgical masks (22.2% correct). Favorable agreement was high across the four attitude statements. Heavy workload (57.8%), lack of training (55.0%), lack of supervision (47.8%), and lack of supplies (45.6%) were the most frequently selected barriers (Table 5). Internal consistency was low for knowledge (alpha = 0.394), practice (0.391), attitude (0.418), and barrier items (0.460).

Table 5.

Knowledge, attitudes, and perceived IPC barriers.

Domain	Item/category	Correct or selected n (%)	Correct answer
Knowledge item	Sterilization eliminates all microorganisms including spores	160 (88.9)	True
	Alcohol hand rub is effective when hands are visibly soiled	57 (31.7)	False
	Hepatitis B can be transmitted during dental procedures	172 (95.6)	True
	Surgical masks provide full airborne-pathogen protection	40 (22.2)	False
	Biological indicators are the gold standard for sterilization monitoring	147 (81.7)	True
Knowledge category	Poor (<60%)	33 (18.3)
	Moderate (60–79%)	86 (47.8)
	Good (80–100%)	61 (33.9)
Attitude: strongly agree/agree	IPC is essential to prevent systemic disease transmission	176 (97.8)
	Patients are concerned about IPC measures	159 (88.3)
	Continuous training improves IPC compliance	155 (86.1)
	Strong supervision increases IPC adherence	157 (87.2)
Barrier selected	Lack of supplies	82 (45.6)
	High cost of materials	64 (35.6)
	Lack of training	99 (55.0)
	Heavy workload	104 (57.8)
	Lack of supervision	86 (47.8)
	Poor infrastructure	57 (31.7)
	Other	21 (11.7)

Knowledge score: mean 3.20 (SD 1.00), median 3 (IQR 3–4), range 0–5. Four-item favorable-attitude score: mean 13.64 (SD 1.79), median 14 (IQR 12–15), range 7–16. Barrier count: n = 179, mean 2.87 (SD 1.60), median 3 (IQR 1–4), range 1–7. Internal consistency was low: knowledge alpha = 0.394; practice alpha = 0.391; attitude alpha = 0.418; barrier alpha = 0.460.

Factors associated with practice score

In univariable analyses, lack of IPC training, absence of a designated officer, inconsistent hand-hygiene supplies, inconsistent PPE supplies, and absence or uncertainty regarding a post-exposure protocol were associated with lower practice scores. These crude associations are presented alongside the adjusted estimates in Table 6 and were not interpreted as independent effects.

Table 6.

Univariable and multivariable robust linear regression of factors associated with IPC practice score.

Variable	Crude beta (95% CI)	p-value	Adjusted beta (95% CI)	p-value
Age, per year	0.041 (−0.046 to 0.127)	0.354	0.067 (−0.042 to 0.176)	0.226
Sex: female	0.268 (−0.412 to 0.948)	0.437	0.239 (−0.695 to 1.173)	0.614
Professional category
General dentist	Reference		Reference
Dental specialist	0.495 (−0.396 to 1.385)	0.274	−0.367 (−1.289 to 0.555)	0.433
Dental therapist	−1.679 (−2.100 to −1.258)	<0.001	−0.542 (−1.741 to 0.657)	0.374
Dental assistant	0.163 (−0.850 to 1.176)	0.751	−0.196 (−1.281 to 0.890)	0.722
Other	−0.679 (−4.946 to 3.588)	0.754	−1.422 (−6.504 to 3.660)	0.581
Education
Diploma	Reference		Not included
Bachelor’s degree	−0.985 (−3.111 to 1.141)	0.362	Not included
Master’s degree	−0.529 (−2.747 to 1.690)	0.639	Not included
Specialist certification	0.733 (−1.622 to 3.088)	0.540	Not included
Other	−2.600 (−4.688 to −0.512)	0.015	Not included
Clinical experience
Less than 1 year	Reference		Reference
1–5 years	−0.106 (−1.265 to 1.053)	0.857	−0.506 (−1.592 to 0.580)	0.359
6–10 years	0.606 (−0.674 to 1.887)	0.351	−0.112 (−1.567 to 1.343)	0.879
More than 10 years	0.316 (−1.334 to 1.966)	0.706	−0.895 (−2.638 to 0.848)	0.312

Facility type
Private clinic	Reference		Reference
Public hospital dental unit	−0.478 (−1.364 to 0.407)	0.288	−0.550 (−1.410 to 0.309)	0.208
NGO clinic	−1.758 (−2.150 to −1.367)	<0.001	−2.332 (−3.574 to −1.090)	<0.001
Teaching clinic	1.842 (1.027 to 2.656)	<0.001	1.223 (−0.127 to 2.573)	0.075
Patients treated/day
Fewer than 10	Reference		Reference
10–19	−0.426 (−1.260 to 0.409)	0.316	−0.894 (−1.673 to −0.114)	0.025
20 or more	0.502 (−0.386 to 1.389)	0.266	−0.302 (−1.156 to 0.552)	0.486
No IPC training	−1.542 (−2.392 to −0.691)	<0.001	−1.437 (−2.469 to −0.406)	0.007
Written guidelines
Yes	Reference		Reference
No	−0.297 (−1.132 to 0.537)	0.483	0.477 (−0.532 to 1.486)	0.352
Not sure	−0.172 (−2.008 to 1.664)	0.853	−0.135 (−1.842 to 1.572)	0.876
No designated IPC officer	−0.896 (−1.654 to −0.138)	0.021	0.018 (−0.858 to 0.895)	0.967
Hand-hygiene supplies
Always	Reference		Reference
Sometimes	−2.044 (−3.294 to −0.794)	0.001	−0.793 (−1.932 to 0.346)	0.171
PPE supplies
Always	Reference		Reference
Sometimes	−2.005 (−2.906 to −1.104)	<0.001	−1.380 (−2.422 to −0.338)	0.010
Rarely	−5.217 (−7.962 to −2.472)	<0.001	−4.994 (−7.023 to −2.964)	<0.001
Knowledge score, per point	0.300 (−0.058 to 0.657)	0.100	0.047 (−0.276 to 0.370)	0.775
Post-exposure protocol
Yes	Reference		Reference
No	−1.634 (−2.568 to −0.700)	0.001	−0.533 (−1.526 to 0.460)	0.291
Not sure	−1.244 (−2.303 to −0.184)	0.022	−0.204 (−1.157 to 0.750)	0.673
PPE selection count, per item	−0.079 (−0.304 to 0.145)	0.486	Not included
Barrier count, per item	−0.171 (−0.399 to 0.058)	0.142	Not included

Outcome: seven-item IPC practice score (0-28). N = 180, except the univariable barrier-count model (n = 179). Robust standard errors were used. Model R-squared = 0.356; adjusted R-squared = 0.256; root MSE = 2.019.

Sparse facility and professional categories should be interpreted cautiously.

CI confidence interval, IPC infection prevention and control, PPE personal protective equipment.

In the adjusted model, participants without recent IPC training had a mean practice score 1.44 points lower than trained participants (adjusted beta −1.44, 95% CI −2.47 to −0.41; p = 0.007). Relative to always-adequate PPE supplies, sometimes-adequate supplies were associated with a 1.38-point lower score (95% CI −2.42 to −0.34; p = 0.010), and rarely-adequate supplies with a 4.99-point lower score (95% CI −7.02 to −2.96; p < 0.001). Participants treating 10–19 patients per day had a 0.89-point lower score than those treating fewer than 10 (95% CI −1.67 to −0.11; p = 0.025); the estimate for 20 or more patients was compatible with no association. Facility type was significant as a joint term (p = 0.0003), but the NGO category contained one participant and the teaching category five; category-specific estimates therefore do not support stable facility-level conclusions. Knowledge score was not independently associated with practice score (adjusted beta 0.05 per point, 95% CI −0.28 to 0.37; p = 0.775). The model R-squared was 0.356, adjusted R-squared 0.256, and root MSE 2.02 (Table 6).

Diagnostics and sensitivity analyses

The practice score departed from normality (Shapiro-Wilk W = 0.894; p < 0.001), and the primary-model residuals also departed from normality (W = 0.940; p < 0.001). Heteroscedasticity was detected (Breusch-Pagan/Cook-Weisberg p = 0.0147), supporting the use of robust standard errors. Multicollinearity was not substantial (mean VIF 1.67; maximum 3.79), and Ramsey RESET did not detect evidence of functional-form misspecification (p = 0.210). Diagnostic plots are shown in Supplementary Fig. S1C–F. Fractional-logit analysis supported the direction of the principal findings for IPC training and PPE supply adequacy. Excluding the two patient-load values above 100 produced materially similar estimates.

Discussion

This study found high self-reported adherence to several routine dental IPC practices, but the overall picture was not uniformly reassuring. Important weaknesses remained in respiratory protection for aerosol procedures, biological monitoring of sterilization, complete occupational vaccination, availability of post-exposure protocols, and dependable PPE supply. In the adjusted analysis, lack of recent IPC training and intermittent or rare PPE availability were associated with meaningfully lower practice scores. These findings shift the interpretation away from a generalized claim of poor practice and toward a more specific conclusion: routine behaviors were commonly reported, while system-dependent and technically demanding components of IPC remained incomplete.

The strong reported performance for hand hygiene, glove use, glove changing, instrument packaging, surface disinfection, and sharps disposal contrasts with the low frequencies shown in the originally submitted tables, which resulted from reversed category presentation. The corrected pattern is more consistent with a dental-provider survey in Jordan that reported generally high use of basic barriers but remaining variation in other infection-control procedures [7]. It is also compatible with the broader IPC literature showing that compliance differs by task and context rather than forming a single uniform behavior [6]. Because these measures were self-reported, ceiling effects and social-desirability bias may have inflated adherence, particularly for visible or professionally expected behaviors. Direct observation and sterilization records are, therefore, needed before treating the high composite score as equivalent to objectively verified safe practice.

Respiratory protection showed a different pattern. Most participants selected surgical masks for aerosol procedures, but only about one-fifth selected an N95 respirator, and fewer than half selected face shields or protective eyewear. CDC guidance distinguishes surgical masks from respiratory protection and recommends selecting PPE according to anticipated exposure and transmission route [3, 4]. The knowledge item on full airborne protection from surgical masks was also frequently answered incorrectly. Together, these findings point to a specific knowledge-practice gap around aerosols rather than a general lack of PPE use. A plausible explanation is that surgical masks and gowns are more routinely available and familiar, whereas respirators require procurement, fit considerations, and risk-based protocols. Facility programs should therefore combine respiratory-protection training with reliable access and locally appropriate risk assessment, rather than relying on information campaigns alone.

Sterilization practices similarly combined broad access to autoclaving with limited use of biological indicators. CDC guidance treats cleaning, packaging, sterilization, monitoring, storage, and documentation as a connected reprocessing system, not as interchangeable activities [3, 4]. Mechanical or chemical monitoring can confirm exposure to process parameters, but biological monitoring provides direct evidence of microbial killing. The observed pattern may reflect cost, procurement difficulty, limited laboratory support, or insufficient assignment of reprocessing responsibility. The practical implication is that supplying autoclaves without routine quality assurance is incomplete; facilities need written workflows, trained responsible staff, indicator schedules, documentation, and corrective procedures for failed cycles.

Recent IPC training remained independently associated with a higher practice score after adjustment. This agrees with systematic-review evidence that training and knowledge are relevant but operate alongside institutional, behavioral, and resource conditions [6]. CDC dental guidance recommends training on initial assignment, when tasks change, and at least annually [4]. In the present study, knowledge score itself was not independently associated with practice, and its confidence interval was compatible with both small positive and negative effects. The earlier claim that greater knowledge predicted worse practice was therefore unsupported. The corrected result is more plausibly interpreted as evidence that a five-item knowledge quiz cannot capture the competencies, routines, resources, and organizational reinforcement required for implementation. Competency-based training should consequently include demonstration, feedback, and facility follow-up rather than knowledge testing alone.

PPE availability showed the largest and clearest adjusted gradient. Participants reporting supplies as sometimes adequate scored lower than those with always-adequate supplies, and the deficit was substantially larger when supplies were rarely adequate. Studies in Tanzania and Ethiopia have similarly described supplies, infrastructure, training, workload, and organizational support as determinants of IPC implementation [9–11]. The dose pattern in this study strengthens the interpretation that resource reliability matters, although cross-sectional data cannot establish temporal direction and self-reported supply adequacy may partly reflect respondents’ overall perceptions of facility quality. Even with that uncertainty, the association identifies a directly actionable management problem: procurement should be monitored by stock availability and interruption frequency, not merely by whether PPE is nominally listed at the facility.

The patient-load finding was non-linear. Scores were lower among participants treating 10–19 patients per day than among those treating fewer than 10, while the estimate for 20 or more patients was not distinguishable from no difference. This does not support a simple conclusion that increasing workload consistently reduces or improves IPC. The category may combine providers working in facilities with different staffing, resources, service types, or reporting practices, and two extreme values further illustrate measurement uncertainty. The materially similar restricted analysis reduces concern that those two values drove the main coefficients, but residual confounding remains likely. Future studies should measure patient flow alongside procedure type, appointment duration, staffing ratios, and directly observed opportunities for IPC.

Nearly one-third of participants reported a needle-stick injury in the previous year, and one-quarter of those exposed did not report it. A global meta-analysis has documented substantial occupational needle-stick exposure among healthcare workers [13], while CDC dental guidance emphasizes injury prevention, immediate reporting, medical evaluation, hepatitis B vaccination, and post-exposure follow-up [4]. Incomplete protocol availability and incomplete vaccination coverage therefore matter even in a sample reporting high routine practice. The finding supports establishing a standardized referral pathway for exposure assessment and confidential follow-up, with facility registers used for prevention learning rather than blame.

Hepatitis B vaccination coverage was higher than in some health-workforce surveys but remained incomplete for an occupationally exposed group; a nationwide Nigerian survey similarly documented gaps between risk awareness and completed vaccination [14]. Vaccination was not retained as an independent predictor in the corrected primary model, so it should not be interpreted as a determinant of practice. Its importance is clinical and programmatic: vaccination directly protects workers from an established occupational hazard and should be provided through employment or professional-registration systems, accompanied by documentation and post-vaccination follow-up according to applicable policy [4].

Facility-type variation should be interpreted conservatively. Although the joint term was associated with practice score, one participant represented the NGO category and five represented teaching clinics. Such sparse cells can produce unstable estimates that appear large because they reflect individual observations rather than reproducible facility effects. The box plot therefore serves as a descriptive display only (Supplementary Fig. S1C). It does not justify claims that NGO or teaching facilities perform better or worse. A future stratified facility survey with verified denominators and adequate numbers in every facility category is required to assess contextual variation.

Strengths and limitations

The study assessed multiple components of dental IPC in a setting with limited published evidence and retained both item-level responses and a transparently rescored composite outcome. Reanalysis from the original Excel export, preservation of raw fields, explicit answer keys, presentation of crude before adjusted models, robust standard errors, diagnostics, and sensitivity analyses improved transparency. Reporting both high-frequency routine practices and weaker system-dependent components provides a more useful account than a single good-versus-poor label.

Several limitations materially affect interpretation. First, the absence of a comprehensive workforce and facility registry meant that the eligible population, number approached, and response rate could not be verified. Selection and non-response bias cannot therefore be quantified, and the findings should not be generalized to all dental professionals or facilities in Somaliland. Second, participant-level city was absent from the analytical dataset. Although fieldwork was described across four areas, regional variation could not be evaluated. Third, practices, vaccination, supplies, and injuries were self-reported. Recall and social-desirability bias could inflate adherence or understate occupational incidents, affecting outcome validity.

Fourth, the cross-sectional design establishes neither temporality nor causality; training and supply conditions were associated with current reported practice but cannot be assumed to have produced it. Fifth, the five knowledge questions and heterogeneous practice, attitude, and barrier items had low internal consistency. The scores should be interpreted as concise descriptive indices rather than validated unidimensional scales. Sixth, several professional and facility categories were very small, producing unstable estimates and limiting precision. The multivariable model also involved numerous parameters relative to the sample size, so residual and unmeasured confounding remain possible. Seventh, multiple exploratory comparisons increase the possibility of chance findings. Finally, no direct observation, sterilization log review, microbiological testing, or facility audit was performed; the study therefore measures reported practice rather than verified IPC performance.

Implications for policy, practice, and future research

Immediate facility action should focus on the gaps identified by the item-level data. Dental facilities should assign a trained IPC focal person, maintain written procedures, provide recurrent competency-based training, and use simple monthly checklists covering PPE stocks, hand-hygiene supplies, sterilizer cycles, indicator results, sharps injuries, vaccination, and post-exposure referrals [3, 4, 15]. Because PPE availability displayed a graded association with practice, procurement systems should establish minimum stock levels, reorder points, and interruption reporting for gloves, masks, respirators, eye protection, face shields, and gowns.

Sterilization quality assurance requires particular attention: every facility using reusable instruments should document cleaning, packaging, cycle parameters, chemical indicators, scheduled biological monitoring, storage, and corrective action. Occupational health programs should offer hepatitis B vaccination and a confidential pathway for immediate exposure assessment. Health authorities, dental professional bodies, and training institutions can support these actions through a standardized dental IPC checklist, supportive supervision, continuing-professional-development requirements, and aggregated reporting of supply interruptions and occupational exposures. Audit and feedback generally produce modest improvements in professional practice and are most useful when repeated, actionable, and linked to clear targets [16].

Future research should begin with a verified census of dental facilities and personnel so that sampling fractions and response rates can be calculated. A stratified facility study should recruit adequate numbers by city and facility type, combine questionnaires with structured observation and record review, and use a validated dental IPC instrument. Repeated audits or a prospective before-and-after implementation study could then evaluate whether dependable supplies, competency-based training, and feedback improve objectively measured practice. Qualitative work with practitioners and managers would help explain procurement interruptions, non-reporting of injuries, and barriers to biological monitoring.

Conclusion

Dental healthcare professionals in the surveyed Somaliland facilities reported high adherence to several routine IPC practices, while important gaps persisted in respiratory protection, sterilization monitoring, occupational vaccination, post-exposure systems, and reliable PPE supply. Recent training and dependable PPE availability were associated with higher practice scores, but causal interpretation is not possible. The unknown response rate, self-reported outcomes, absent city linkage, and sparse subgroups limit accuracy and generalizability. Facility-level programs should prioritize competency-based training, uninterrupted supplies, sterilization quality assurance, vaccination, post-exposure management, and auditable IPC procedures, followed by prospective evaluation using verified sampling frames and direct observation.

Author contributions

DKY conceived the study, developed the methodology, coordinated data collection, conducted the statistical analysis, interpreted the findings, and drafted the manuscript. AAH contributed to data validation, interpretation of the corrected findings, and critical revision of the manuscript. Both authors reviewed and approved the final manuscript and agree to be accountable for the work.

Data availability

The de-identified dataset and analysis code are available from the corresponding author on reasonable request, subject to institutional approval and protection of participant confidentiality.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41405-026-00483-4.