Hand Hygiene Behaviors Associated With Peritoneal Dialysis-Hand Hygiene Behaviors Associated With Peritoneal Dialysis-Related Infection Related Infection

Background : Hand hygiene (HH) is an effective strategy to prevent infections. This study aimed to explore the association between HH behavior among peritoneal dialysis (PD) patients or caregivers and PD-related infection. Method : A single-center, cross-sectional study was conducted among PD patients and primary caregivers of totally-dependent PD patients. Data were collected through a participant information form and three different questionnaires about knowledge of HH, facilities for HH, and HH behavior. Documentation of PD-related infections were obtained from patients ’ medical records during the last 12 months. Descriptive statistics and binary logistic regression were used to analyze the data. Results : A total of 254 participants, including 181 PD patients (71%) and 73 caregivers (29%), were studied. The overall mean score for HH behavior was at a ‘ good ’ level (3.7 ± 0.3). Nearly half of the participants (46%) had a PD-related infection. In the binary logistic regression analysis, ‘ improper ’ hand and ﬁ ngernail hygiene (OR 1.72, 95% CI: 1.04 e 2.84), ‘ improper ’ hand-washing steps (OR 1.68, 95% CI: 1.02 e 2.80), and ‘ improper ’ hand-washing during PD procedures (OR 2.32, 95% CI: 1.20 e 4.50) were signi ﬁ cantly associated with PD-related infection. Conclusion : ‘ Improper ’ HH behavior is associated with increased risk of PD-related infection. Our ﬁ ndings serve as input to develop ef ﬁ cient training and retraining programs for Thai PD patients and primary caregivers to improve their HH behavior and reduce risks of PD-related infection. Further investigation of training and retraining programs and of hand-washing techniques for PD is warranted.


Introduction
P eritoneal dialysis (PD) is a home-based renal replacement therapy for end-stage kidney disease (ESKD) patients [1,2]. The incidence of ESKD has been increasing worldwide. According to the Thai Renal Replacement Therapy (TRT) registry, the number of PD patients in Thailand has expanded nearly seven-fold from 5133 to 34,467 from 2009 to 2020 [3]. This is partly due to the PD-first policy adopted by the Thai government in 2008 as part of the universal health coverage scheme (UCS). As a result, continuous ambulatory PD (CAPD) is provided as the first dialysis modality for kidney replacement therapy in ESKD patients (unless contraindicated) [4].
To remove waste substances, a cleansing solution (i.e., dialysate or PD fluid) is introduced into the peritoneal cavity through a PD catheter. The PD catheter consists of the Tenckhoff catheter and the extension part called 'transfer set.' The peritoneal membrane works as a filter. The excessive substance molecules, uremic toxins, and fluid are removed through the semipermeable peritoneal membrane according to concentration gradient until reaching equilibrium. These exchanges can be done manually, three to four times per day by CAPD, or by automated PD (APD) machine, depending on clinical indication, patient convenience, and accessibility. In either technique, the patients need to perform PD bag exchanges, make a cycler connection-disconnection, and apply a daily exit-site dressing. However, some patients cannot perform such procedures by themselves. Thus, they need support from their caregiver [2].
One of the common PD complications is PD-related infection, including PD-related peritonitis and exitsite infection (ESI). These infections are an indicator of technique failure, which may require a switch of mode of renal replacement therapy to hemodialysis. This switch could increase risk of mortality [5]. The mortality rate in PD-related peritonitis was reported to be 19.8% [6]. The major mechanism that causes PDrelated infection is touch contamination by unwashed or poor-hygiene hands. Gram-positive bacteria are the most common pathogenic bacteria leading to PDrelated infection. Positive bacterial cultures (Staphylococcus) have been found from patients' hands [7]. Because patients or caregivers need to perform multiple procedures each day, strict adherence to 'proper' hand hygiene (HH) without contamination is required to prevent PD-related infection [8]. However, different HH techniques have been used to train PD patients before therapy.
HH is a simple but effective strategy to prevent healthcare-related infections, as it helps reduce the spread of floral and multi-resistant organisms [9]. However, most patients and caregivers neglect HH. Dong and Chen [10] reported that over half (52%) of PD patients and caregivers had 'improper' HH. Ye et al. [7] explored a six-step hand-washing protocol based on the hospital hygienic standards in 86 Chinese patients with PD. They found that only 26% of patients adhered to the hand-washing protocol. In addition, Mawar et al. [11] observed the CAPD exchange procedure in 30 Indian patients undergoing CAPD at their home. The study found that poor compliance in performing the CAPD procedure was significantly associated with a higher incidence of PD-related peritonitis (p < 0.03) [11]. Improperly performed steps included not wearing a sanitary mask, not flushing before filling, and not washing hands [11].
In Thailand, the number of studies related to HH behavior and PD-related infection is very limited. One such study is a case report. In this report, the PD patient was diagnosed with catheter-related peritonitis from fungal infection, Aureobasidium pullulans [12]. The authors reported that 'improper' HH and fingernail care were potential causes of this infection [12]. A descriptive study examined selfcare behaviors such as diet, exercise, and PD exchange behaviors in 21 PD patients. Data were collected by semi-structured interview and observation [13]. The results indicate that 39% had 'improper' behaviors, including omitting one or more hand-washing steps, unsterile PD bag exchange, and 'improper' exit-site dressing [13]. Another descriptive study explored infection rate and factors associated with PD-related infection in 22 CAPD patients [14]. That study found that uncovering the PD wound during bathing and incorrect hand-washing were factors related to PDrelated infection [14].
Although previous studies reported an association between HH and PD-related infection, all of them were conducted with different definitions and measurement of HH under different research methods and among small sample sizes. Thus, little is known about the association between HH behavior and PDrelated infection in the Thai population. To our knowledge, no previous study has explored comprehensive HH behavior that covers hand and fingernail hygiene, preparation before handwashing, hand-washing before, during, and after PD procedures, and hand-washing steps. The differences in terms of HH behavior definition, protocol, and measurement limit the generalizability of the results of previous studies. Currently, standard training and retraining programs have not been implemented in Thailand. Most PD clinics develop their own training programs regarding HH behavior and hand-washing techniques to train their PD patients and caregivers. HH knowledge and hygieneenabling facilities required for PD procedures at home may be different from clinic to clinic. Understanding HH behavior and PD-related infection in Thailand are essential in the development of standard training and retraining programs. Therefore, this study explored the association between HH behavior and PD-related infection among PD patients and caregivers in a Thai population using questionnaires based on recommendations by WHO and Thai nursing practice guidelines for hemodialysis and PD [15,16]. HH knowledge and HH facilities for PD are also assessed in this study to gain more understanding of HH knowledge level and adequacy of HH facilities. The findings from this study can be used to implement training and retraining programs, improve 'proper' HH behavior, and prevent PD-related infection in the Thai PD population.

Study design and participants
This is a cross-sectional study. The participants were recruited from a single PD clinic in Bangkok, Thailand from April to June 2020. PD patients, or caregivers who attended the clinic, were screened for eligibility. Eligible participants were age 18 years or older, had been performing PD bag exchange and exit-site dressing for more than one month, could complete the study questionnaires, and gave informed consent. Disabled or bedridden patients, patients with a cirrhosis Child score of Class C or severe co-morbidities (e.g., advanced cancer, pregnancy, hybrid renal replacement therapy) were excluded from the study.

Sample size calculation
The desired sample size was calculated using a formula based on an estimate proportion of the population that demonstrated proper HH technique [17]. The total number of PD patients (N) at the clinic was 888. Proportion (p) was set at 0.64 because a previous study found 64% of PD patients had 'proper' hand-washing [14]. Error (d) was set at 0.05, alpha (a) was set at 0.05, and Z was set at 1.96. A minimum sample size of 254 participants was needed to detect statistical significance.

Operational definition
"HH behavior" in this study is defined as actions that PD patients and caregivers perform in seven dimensions including: 1) Hand and fingernail hygiene; 2) Hand preparation before washing; 3) Wearing a sanitary mask before hand-washing; 4) Hand-washing steps; 5) Hand rubbing with alcohol gel; 6) Hand-washing during PD procedures; and 7) Hand-washing after PD procedures. HH behavior was measured by the HH Behavior for PD Procedure Questionnaire. Each dimension was categorized as either 'proper' or 'improper' HH behavior. However, proper and improper HH behaviors are not universally defined. In this study, we defined 'proper' HH behavior as every-time accurate behavior, and 'improper' HH behavior as not every-time accurate behavior.

Part 1: participant information form
This form was developed by the researchers. It was used to collect information about age, gender, level of education, occupation, family income, income sufficiency, and duration of PD procedures of the participants who were PD patients or caregivers of totally-dependent PD patients. Data regarding peritonitis and ESI were obtained from patients' medical records during the last 12 months, and summarized as "having infection" or "no infection". If the participants were caregivers, permission to review the information from patients' medical records was also obtained from the PD patients.
2.4.2. Part 2: questionnaire to measure knowledge of HH for PD The questionnaire to measure knowledge of HH for PD was developed by the researchers based on recommendations for HH by WHO [15] and Thai nursing practice guidelines for hemodialysis and PD [16]. Knowledge of HH included the importance of HH, hand-washing steps, hand-washing technique, materials, and hand-washing duration for PD. The questionnaire consisted of 12 items. Each item could be responded to by one of three choices: "Yes," "No," and "Not Sure." The knowledge score had a potential range of 0e12 points. It was divided into three categories: 'good' (>75% of total score), 'fair' (60e75% of total score), and 'poor' (<60% of total score) [18]. The questionnaire was checked for content validity and language suitability by five qualified experts. Their check yielded a content validity index (CVI) value of 0.97. It was externally validated in 30 PD patients and caregivers (not participating in the main study) with a test-retest reliability coefficient of 0.86.

Part 3: questionnaire to measure facilities for HH for PD
The questionnaire to measure the facilities for HH for PD was developed by the researchers based on the literature review [15,16] and adapted to the context in Thailand. This questionnaire was used to assess the use of appropriate equipment (e.g., PD basin and faucet), and materials (e.g. clean water, paper towels/towels, soap, and alcohol gel) for HH at home. The questionnaire consists of ten items, and the response options use a 5-level Likert scale consisting of "Every time (5 points)," "Almost every time (4 points)," "Occasionally (3 points)," "Almost never (2 points)," and "Never (1 point)." The total response score was calculated by taking the sum of the ten responses. Thus, the total score range was from 10 to 50 points. Scores were grouped into three categories 'good' (38e50), 'fair' (24e37), and 'poor' (10e23) facilities for HH [19]. The questionnaire was checked for content validity and language suitability by five qualified experts, yielding a CVI value of 0.98. The questionnaire was externally validated in 30 PD patients and caregivers (not participating in the main study) and in 254 participants, yielding Cronbach's alpha coefficients of 0.80 and 0.72, respectively.

Part 4: questionnaire to measure HH behavior for PD procedures
The questionnaire to measure HH behavior for PD procedures was developed by the researchers based on WHO guidelines [15] and Thai nursing practice guidelines for hemodialysis and PD [16]. The questionnaire consists of seven dimensions, including hand and fingernail hygiene (2 items), hand preparation before washing (2 items), wearing a sanitary mask before hand-washing (1 item), hand-washing steps (12 items), hand rubbing with alcohol gel (1 item), hand-washing during PD procedures (3 items), and hand-washing after PD procedures (1 item). Response to each item was rated on a 4-point scale as follows: "Never," "Sometimes," "Frequently," and "Every Time." Potential scores ranged from 1 to 4 points. A higher score indicated better HH behavior. The overall scores were grouped into three categories: 'good' (3.1e4.0), 'fair' (2.1e3.0), and 'poor' (1.0e2.0) [19]. Considered by dimension, when the participants reported "every time" (score ¼ 4 points) on all items of each dimension, that dimension was coded as 'proper.' For any response of fewer than 4 points, the dimension was coded as 'improper.' The questionnaire was checked for content validity and language suitability by five qualified experts, yielding a CVI with a value of 0.91. The questionnaire was externally validated in 30 PD patients and caregivers (not participating in the main study) and in 254 participants, yielding Cronbach's alpha coefficients of 0.85 and 0.76, respectively.

Data analysis
Data were analyzed using SPSS version 18.0 (SPSS Inc., Chicago, II., USA). Categorical data were described by frequencies and percentages. Normally-distributed continuous variables were described as mean ± standard deviation (SD). Nonnormally-distributed continuous variables were illustrated as median (IQR). We used Pearson's product moment correlation to identify the associations between age, dialysis vintage, family income, HH knowledge, HH facilities and HH behavior. Spearman's Rho correlation was used to validate the association between level of education and HH behavior. Point-biserial correlation was used to identify the associations between type of participant, gender, employment status, sufficiency of family income, and HH behavior. In addition, the association between HH behavior and PD-related infection was tested by binary logistic regression. All p-value calculations were two-sided, and deemed statistically significant at a ¼ 0.05.

Ethical considerations
The protocol for this study was approved by the Institutional Review Board, Faculty of Nursing, Mahidol University, Thailand (COA No. IRB-NS2020/556.3003).

Participant characteristics
A total of 254 participants, including 181 PD patients (71%) and 73 primary caregivers (29%) were enrolled and completed the study questionnaires. More than half of the participants were female (52%). The mean age of participants was 54.1 ± 12.8 years. Half of the participants had completed elementary school (50%). Nearly three out of five participants were currently employed (58%), had a monthly family income of 10,000 baht ($300) or less (59%). Under two-thirds of participants reported having sufficient income (64%). The mean duration of PD therapy was 34.0 ± 29.9 months.

HH knowledge and facilities for HH
The HH knowledge score (maximum potential of 12 points) was at a 'fair' level (mean 8.0 ± 2.0, range of 1e12). The facilities for HH score (maximum potential of 50 points) was at a 'good' level (mean 46.2 ± 4.6, range 30e50) ( Table 1). HH knowledge (r ¼ 0.217, p ¼ 0.001) and facilities for HH (r ¼ 0.298, p < 0.001) were positively correlated with HH behavior (Table 1).

HH behavior and PD-related infection
The overall mean score for HH behavior when performing PD procedures was 3.7 (SD ¼ 0.3) indicating a 'good' level (Table 1). However, when considering HH behavior for PD for each dimension, 54% of participants had 'improper' hand and fingernail hygiene, 59% had 'improper' handwashing steps, and 18% had 'improper' handwashing during PD procedures. PD-related infection was found in 46% of participants. Of those with PD-related infection, 47% had ESI, 28% had PDrelated peritonitis, and 25% had concomitant ESI and peritonitis. The most common causative organism was gram-positive bacteria, which was found in 46% of PD-related peritonitis and 62% of ESI cases.

Discussion
This study investigated the association of HH behavior in PD patients and caregivers and PDrelated infections in Thailand. Among the seven dimensions of HH behavior, hand-washing steps and hand/fingernail hygiene were the two most common 'improper' behaviors. PD-related infection was significantly associated with 'improper' hand and fingernail hygiene, 'improper' hand-washing steps, and 'improper' hand-washing during PD procedures.
HH was neglected by some participants, and the hand-washing steps tended to be incomplete. The hand-washing steps that were lacking included: interlacing the backs of fingers to opposing palms; rotational rubbing of thumb clasped in the opposite palm; and rotational rubbing, backward and forward, with clasped fingers of the hand in the opposite palm, and vice versa; and practicing an adequate duration of hand-washing (i.e., at least 40 s). In terms of 'improper' hand and fingernail hygiene (54%), some participants may not know that washing hands 'properly' without trimming the nails may leave germs at fingertips and under the nails. Key factors influencing HH behavior include having undergone a training program, recognition of the importance of HH, receiving repeated instructions, and reaffirmation of HH techniques by PD nurses at the clinic. Although the frequency of washing hands during the procedure was 'good,' there was still a lack of connection between knowledge and application in actual practice. As for antiseptic techniques, most participants thought that washing hands with liquid soap and clean water was sufficient to eliminate germs on their hands. These findings are similar to those of a study by Ye et al. [7], which found that 74% of participants did not follow all six standard steps of hand-washing, and 26% did not use soap to clean their hands. Moreover, beyond the hand-washing techniques, longer fingernails also affect HH as they lead to inadequate hand-washing and can harbor dirt and bacteria. Thus, longer fingernails potentially predispose the patient to infection. All participants need to wash their hands before, during, and after the procedures. However, 18% of participants missed the steps of hand-washing during the procedures, because they thought that they were not necessary. Also, some participants may have found correct HH behavior was inconvenient due to the lack of availability of HH facilities.
The results show that HH knowledge for PD was positively correlated with HH behavior (r ¼ 0.217, p ¼ 0.001). Among HH knowledge scores, the item that yielded the highest score was "washing hands when dirt was on their hands," while the lowest was "reusing alcohol and liquid soap containers." Similar to our study, a study conducted among inpatients found that high knowledge scores corresponded to high HH scores [20]. Although every participant had been educated on HH knowledge throughout the training, some knowledge eroded over time as the average duration of dialysis was approximately 2 years and 8 months. Similarly, the study by Suanpoot [14] found that caregivers of PD patients with peritonitis mostly had misunderstood concepts of HH, i.e., they believed that using disinfectants alone was enough, even without 'proper' hand-washing.
The facilities for HH also had a positive relationship with HH behavior (r ¼ 0.298, p < 0.001). Although this study found that the facilities for HH had a 'good' mean score, the most common deficiency was lack of a 'proper' basin to practice HH, primarily due to the lower income of the family.
Installing a 'proper' basin for HH involves cost for the basin, related equipment, installation, and clean water access. Similar to the observational study of HH among Thai PD patients and caregivers who experienced infectious complications, the participants who could not access tap water poured filtered water from a tumbler or a bottle instead [13]. Another study found that 59% of patients with peritonitis did not have a washing basin [14]. A study conducted among professional nurses also found that access to HH facilities affected the HH behavior [21]. Therefore, training and retraining programs should emphasize availability of HH facilities to improve HH behavior in the Thai PD population.
The participants who had 'improper' hand and fingernail hygiene, 'improper' hand-washing steps, and 'improper' hand-washing during PD procedures were 1.72 times (OR 1.72, 95% CI: 1.04e2.84, p ¼ 0.034), 1.68 times (OR 1.68, 95% CI: 1.02e2.80, p ¼ 0.044), and 2.32 times more likely to have PDrelated infection (OR 2.32, 95% CI: 1.20e4.50, p ¼ 0.012), respectively. These results suggest that hand and fingernail hygiene, hand-washing steps, and hand-washing during PD procedures are important HH behaviors to prevent PD-related infection. However, this study found that some participants did not rub their hands with alcohol gel during PD procedures. According to a HH study, alcohol-based hand rubs can reduce bacteria more effectively than hand-washing with plain soap (3.2e5.8 log 10 colony-forming unit (CFU) and 0.6e1.1 log 10 CFU, respectively) [22]. The presence of environmental microorganisms or skin flora, when combined with 'improper' hand-washing could cause PD-related infection. Similarly, the study of Mawar et al. [11] found that poor compliance in PD procedures, e.g., not wearing a sanitary mask, not checking on PD bags, or poor HH techniques, were all associated with a higher incidence of PD-related peritonitis. The previous case report in Thailand also found that 'improper' HH and fingernail care could cause PD-related fungal infection [12]. Also, the study of Dong and Chen [10] found that among patients with peritonitis, half had 'improper' HH before performing the PD procedure. From our study, hand-washing during PD procedures, hand and fingernail hygiene, and handwashing steps are crucial. These three aspects of HH should be emphasized in the training and retraining programs for Thai PD patients and their caregivers. Other characteristics (e.g., age, gender, education level, duration of dialysis, employment status, family income, and HH retraining) were not associated with HH behavior for PD in the study. Fung et al. [23] also found that dialysis duration and education level were not associated with HH behavior. However, higher hand-washing error scores were found in patients with more extended dialysis periods [7]. In contrast to our study, another study found that younger patients tended to have poor adherence to the therapeutic dialysis regimen [24].
This study has some limitations. First, this was a cross-sectional study. Thus, we could not test for causality between the independent and dependent variables. Second, the study included only participants who visited a single PD center in Bangkok, Thailand. Third, the time of data collection was not immediately after the infection occurred. Therefore, some patients might have changed their HH behavior at the time we collected the data since the infection occurred. The collected HH behavior data might not accurately represent HH behavior at the time of infection. Finally, the data could be subject to selection bias as urban PD centers may have better facilities than those in rural areas.

Conclusions
Although HH is recommended in the standard guidelines, patients and caregivers may minimize the significance of HH. They may neglect handwashing during PD procedures, hand-washing steps, and hand and fingernail hygiene, which are three aspects of HH associated with risk of PDrelated infection. Efficient training and retraining among PD patients and primary caregivers to improve HH behavior are crucial and should be included in a PD quality improvement program. Additionally, further research should address effective training and retraining programs and hand-washing techniques to prevent PD-related infection in this population group.