INTRODUCTION
Urinary incontinence (UI) is a prevalent, costly condition with personal and social impacts. It affects people of all ages and can occur at any stage of life.1 Globally, UI is underreported, underdiagnosed, and undertreated.1 This health inequity may partially relate to gaps in healthcare providers' knowledge about UI, and/or gaps in education. Nurses and nursing assistants (NAs) represent the largest group of healthcare providers in most countries and are likely to care for people with UI and continence care needs on a regular basis.
According to Paterson,2 all RNs should be equipped with knowledge and understanding about the physiological, psychological, and social aspects of incontinence (urinary and fecal) and be aware of strategies to prevent incontinence and promote continence. Based on 2 studies about education for nurses, it is doubtful that RNs are educationally prepared to address the care needs of people with UI or to promote continence.3,4 A survey of 294 (81%) undergraduate education programs for healthcare practitioners in the UK found that the mean number of hours of education on incontinence in courses on adult nursing was 7.3 hours (SD = 4.8).3 A similar picture emerged from of a survey of undergraduate and graduate nursing students from 46 states in the United States, wherein the average education content about incontinence in nursing school curricula in 1994 was 2.14 hours (SD = 1.72).4
We hypothesized that gaps in nurses' and NAs' knowledge about UI may contribute to poor assessment and management. This hypothesis is supported by research identifying inaccurate beliefs about UI as a normal part of aging and not treatable.5-9 It follows that this belief could negatively affect care providers' clinical decision-making, limit the choices they present to patients with incontinence, and negatively impact the quality and effectiveness of care they provide. Similarly, negative attitudes about UI or toward people with UI may lead to poor care delivery or lack of follow-up care.10 If nurses or NAs harbor inaccurate beliefs about UI or misinterpret patients' efforts to self-manage and conceal this condition, caregiving interactions are likely to be characterized by tension between caregiver and patient.11,12
Education about UI should improve nurses' and NAs' knowledge about, and attitudes toward, UI that translates into effective continence care. According to a systematic review of the effect of education for staff about interventions for UI and fecal incontinence (FI),13 current evidence is limited to 1 controlled trial13 and 1 uncontrolled trial.14 There is need to build on the findings of this systematic review13 to identify and appraise the complete body of research about the effect of education about UI on nurses' and NAs' knowledge, attitudes, practices, and patient outcomes. The purpose of this systematic review is to describe, critique, and summarize research about the effects of education about UI on nurses' and NAs' knowledge about UI, attitudes toward UI, continence care practices, and patient outcomes.
METHODS
We completed a systematic review using PRISMA guidelines.15 Eligible studies were identified through searching PsycINFO, MEDLINE, CINAHL, Web of Science, and Cochrane databases using the following terms: "Nurses or Nursing Aides," "education," "training program," "treatment outcome," "staff development," "staff training," "education program," "workshop," "education package," "intervention," "coursework," "coach," and "coaching, incontinence, urinary incontinence, bladder incontinence." Other publications were identified from the reference lists of relevant publications and from a search of gray literature. The searches were undertaken by E.C. with advice and support from a professional healthcare librarian. The studies were limited to publications between January 1990 and October 2018. Searches were restricted to English-language articles and human studies.
Studies were included if they were randomized controlled trials (RCTs) or nonrandomized or quasirandomized trials with a UI educational intervention that reported pre- and postquantitative data about the effects of education programs designed to evaluate or improve nurses' and/or NAs' knowledge, attitudes toward, and practices about caring for people with UI. We also sought research evaluating the effect of educating on pertinent continence outcomes.
Knowledge was operationally defined as the comprehension and understanding of acquired facts or information about the causes and management of UI, typically requiring a "yes/no" or levels of agreement response. Attitude was operationally defined as an emotional reaction or predisposition about caring for a person with UI. Practice was operationally defined as continence care practices, including practices to prevent and manage UI, and adhere to best practice recommendations. Patient outcomes of interest were knowledge, frequency or severity of UI, presence or absence of UI, and health-related quality of life.
One reviewer (J.O.) extracted data from all the trials and a second reviewer (E.C.) conducted random reliability checks. Data were extracted on the aim of the trial, study design, sample and setting, methods, nature of the intervention, and findings. The Mixed Methods Appraisal Tool was used to evaluate the quality of included studies.16 Each item was rated as "yes" if it clearly met criteria, "no" if it did not, "unclear" if it could not be determined from the available information, and "not applicable" if the specific quality question did not apply to the study design.
LITERATURE SEARCH
The initial search returned 4249 studies and 79 duplicate elements were excluded. Of the remaining 4170 publications, 4069 were excluded based on title search, leaving 101 publications that were read in full. This process resulted in elimination of 82 studies. The main reason for excluding studies was because the study did not meet the design criteria and/or lacked pre- and postevaluative data about the review outcomes of interest. The results of our literature search are summarized in the Figure.
RESULTS
We reviewed 19 trials in depth that met inclusion/exclusion criteria. Most trials were conducted in nursing homes in the United States. Collectively, these studied enrolled 1301 participants (911 nurses, 235 NAs, and 155 unclear). Nursing assistants were included with nurses in 9 trials,17-25 all of which were conducted in nursing homes. In most cases, participants' qualifications were not reported; however, we identified 5 different cohorts, as described in Table 1.
Most trials evaluated one or more outcomes of interest (knowledge, attitudes, practice, and patient outcomes); none addressed all outcomes. Most of the data collection instruments had face or content validity, as indicated in Table 2. Methods to evaluate patient outcomes varied from relying on patients or staff to rate patients' continence status22,24,26-28 to objectively checking patients' continence status (ie, wet checks or pad weights).19,20,23,25 Five trials reported the reliability and/or validity of the associated data collection instruments.19,24,26,27,29
All 19 trials contained descriptive information about the topics that were addressed in the education intervention, as shown in Table 3. The theoretical basis for the education intervention was described in 3 trials using a diffusion of innovation,20 translation science,32 and group problem-solving approaches theoretical framework.14 The nature of the intervention varied in terms of the duration, delivery mode, educational content, expected learning outcomes, assessment methods, and extent to which the teaching methods accommodated different cultural, literacy, and contextual learning needs, as indicated in Table 4.
Study Quality
Studies differed considerably in terms of their design; 7 trials included a control or comparison group.17,19,24,26,29,30,34 Sample sizes also varied (range 4-176 nurses and 21-64 NAs). Only one trial included a power calculation in order to ensure a sample size sufficient to detect a statistically significant difference in the outcomes of interest. The methodological quality of the studies varied from 25% to 75% for uncontrolled trials, as detailed in Table 5, and 25% to 100% for controlled trials, as indicated in Table 6. Results were not able to be pooled due to methodological heterogeneity; therefore, results were reported descriptively.
EFFECT OF EDUCATION ON KNOWLEDGE
Ten trials evaluated nurses'/NA's pre-/posteducation knowledge about UI: 3 of which were RCTs17,30,34 and 7 were nonrandomized trials,14,18,20-22,31,33 as indicated in Table 7. Considered collectively, findings indicated that nurses and NAs had limited ability to determine UI type and factors that require assessment, interpret clinical data to make a diagnosis, and identify patients who may be suitable for active interventions. All researchers who evaluated the effects of UI education on knowledge reported postintervention improvements, which was statistically significant for some studies and for some items on knowledge questionnaires. For example, Mathis and colleauges21 reported statistically significant improvements in participants' abilities to identify stress, functional and overflow UI. Campbell and colleauges17 reported a slight but significant difference by group and time, F (2,135) = 3.39, P < .05, in knowledge scores, favoring the intervention group in a quasiexperimental trial with repeated measures and a control group in a nursing home setting. Bignell and Getliffe30 found statistically significant improvements in community and district nurses' knowledge about antimuscarinic drugs as a possible treatment for UI in phase 3 (P = .001), the need for a physiotherapy referral (P = .033), and modifying caffeine intake (P = .037). Mention of absorbent products was also significantly reduced (P = .049). The following section details the findings according to the different settings in which they were conducted.
Care Setting
Of the 7 trials that described the effect of education on nurses' and NAs' knowledge in nursing homes, 6 uncontrolled trials reported statistically significant improvements in knowledge for some items on the knowledge questionnaire.14,18,20-22 However, in the only controlled trial in this setting,17 there were no significant differences between groups at 12 weeks.
Two trials were identified that reported the effects of education on community or primary care nurses' knowledge: 1 controlled trial30 and 1 nonrandomized clinical trial.31 Both reported statistically significant improvements in participants' knowledge of UI.
Only one trial that met inclusion criteria occurred in the acute/subacute care setting. Williams and colleagues34 reported improved knowledge scores among 117 hospital-based RNs in the UK after disseminating a clinical handbook about continence care and compared to a control group.
One uncontrolled trial reported improved knowledge following a workshop about male catheterization and general continence care in a sample of 130 UK nurses.33 The practice setting was not reported.
EFFECT OF EDUCATION ON ATTITUDES
Two trials were retrieved that provided quantitative data about the effect of education on nurses' or NAs' emotional reaction or predisposition about caring for a person with UI: 1 was a nonrandomized trial14 and 1 was an RCT.17 Campbell and colleagues17 reported positive attitudes at baseline, at the end of a 12-week educational program about prompted voiding and again 12 weeks later. Collette and associates14 reported significantly improved attitudes immediately after an educational intervention (increase of 5.83%; P = .017), but they noted this improvement was not sustained 9 weeks after the completion of the program (decrease of 3.04%; P = .014).
Considered collectively, findings were mixed. De Gagne and coworkers31 also reported that they evaluated nurses' attitudes about UI; however, closer scrutiny revealed that the items measured beliefs rather than attitudes as defined in this systematic review.
EFFECT OF EDUCATION ON CONTINENCE PRACTICES
Eight trials reported pre-/postdata about the effect of education on nurses'/NAs' UI continence care practices: 3 of which were controlled,26,29,30 as shown in Table 8. Five nonrandomized studies reported posteducation improvements in continence care practices,14,20,22,27,32 but this was not borne out in the 3 controlled trials. These findings are discussed based on care setting.
Care Setting
Three studies were conducted in nursing homes, and none were randomized.14,20,22 Results from all 3 indicated improvements in continence care practices, including an increased ability to plan effective interventions,14 more frequent continence assessments,22 and adoption of policies and procedures to embed a continence program into practice.20
Three trials were set in the community.26,27,30 They quantitatively evaluated nurses' adherence to recommendations to screen/assess and manage UI in community-dwelling people with UI; 2 were RCTs26,30 and 1 was a nonrandomized trial.27 Again, findings were mixed. Sampselle and colleagues27 reported an increase in the frequency of nurses' identification of UI following a 3-year project to improve the initial evaluation and treatment of UI of women attending ambulatory clinics in the United States. Bignell and Getliffe30 also reported positive improvements in the intervention group's continence care practices. Specifically, they found a significant reduction in the number of prescriptions for absorbent pads, as well as an increase in UI monitoring, testing patients' urine, and in treatment planning. By contrast, Cheater and associates26 found that nurses' adherence to assessing and managing UI at 6 months did not differ significantly between a comparison group (education materials alone) or 1 of the 3 experimental groups: (1) an audit and feedback group, (2) an education outreach group, and (3) an audit and feedback with education outreach.
Two studies with pre-/postdata about the effect of education on nurses'/NAs' UI continence care practices were conducted in the acute/subacute care setting: 1 was a before-after study and32 trial and 1 was an RCT.29 The before-after trial was set in a neuroscience inpatient unit for patients following stroke located in the United States.32 The researcher reported that nurses' adoption of an evidence-based bladder protocol about prompted voiding increased 2-fold: the mean adoption rate preintervention was 18.1% and 33.4% postintervention. The RCT was conducted in 12 stroke services in the UK.29 It was designed as a 3-arm, parallel, open, exploratory, pragmatic, cluster RCT. The aim was to determine the effects of implementing a "systematic voiding program" (SVP) compared to SVP combined with facilitation (SVP + F) versus usual care (UC) for the management of UI after stroke in secondary care. The SVP intervention comprised bladder training and pelvic floor muscle training for patients who were cognitively able and prompted voiding for patients with cognitive impairments. Participants receiving SVP + F received support from at least one specialist practitioner whose role was to help staff work together, provide the necessary information and training, maintain motivation, and give feedback and practical help when needed. Participants in both intervention groups had access to online training in bladder scanning as well as face-to-face and web-based theoretical and practical education about the SVP.
Staff adherence was one of several outcomes of interest. The researchers found both groups had similar, but low rates of documentation of patients' voiding times (38.9% in the SVP group and 31.9% in the SVP + F group). Rates of adherence to toileting patients within 30 minutes of their scheduled time were also comparable (54.8% of occasions in the SVP group and 56.0% of occasions in the SVP + F group). Similarly, staff adherence to the requirement to document when they prompted patients to the toilet was comparable (57.9% of occasions in the SVP group and 65.9% of occasions in the SVP + F group). Although patients' catheters were removed in a timelier manner in the SVP + F group, there were no statistically significant differences between the 2 groups.
EFFECT ON PATIENT OUTCOMES
Eleven trials reported the statistical effect of UI education on patient outcomes: 4 were RCTs19,24,26,29 and 7 were nonrandomized studies (Table 9).20,22,23,25,27,28,32 Six studies were conducted in the nursing home setting, 4 of which were nonrandomized studies20,22,23,25 and 2 were RCT.19,24 Data from the uncontrolled trials were mixed. Lekan-Rutledge20 reported a reduction in UI rates among nursing home residents 3 and 6 months after a prompted voiding toileting assistance program that included staff education, staff management, and quality monitoring; however, these differences were not statistically significant. Similarly, Rahman and colleagues22 reported that residents were more continent after nursing home staff had attended a distance coaching course to facilitate the adoption of evidence-based protocols for UI management; outcomes were based on supervisors' opinions. Whilst nursing home residents in the trial by Vinsnes and coworkers25 experienced a significant reduction in the severity of UI (ie, a reduction in the volume of urine loss), the number of pad changes and average postvoid residual volumes remained unchanged. Remsburg and associates23 found that the continence status of most residents remained the same or declined following an education program for staff about UI, its causes, and prompted voiding.
Sackley and coinvestigators24 conducted a phase II exploratory cluster RCT to evaluate the effects of a training program for staff about UI and mobility support compared to standard care. Although the number of patients who reported being continent at 6 weeks increased, the trial was inadequately powered to determine the magnitude of effect produced by the intervention.
The strongest evidence of the effects of education in the nursing home setting derives from a large stepped-wedge RCT conducted in Switzerland. Kohler and collegues19 randomized residents with dementia to the intervention on a stepped basis. The 4-hour education session for RNs and NAs focused on UI and dementia and the management of behavioral and psychological symptoms of dementia. In addition, the researchers convened 6 inpatient care-based case conferences. While UI decreased between baseline and follow-up at 6 months, no significant between-group differences were seen at the study end point (at 14 months for cluster 1, 11 months for cluster 2, 9 months for cluster 3, and 7 months for cluster 4).
Findings from the 3 trials undertaken in ambulatory community or primary care settings also reveal mixed results.26-28 The strongest evidence about the effects on patient outcomes of continence education for nurses working in the community derives from an RCT by Cheater and coworkers,26 who reported that while nocturia and voiding frequency improved in up to half of all patients in 6 months, improvements were similar across all groups. In a nonrandomized trial, Sampselle and colleagues27 found significant increases in women's self-rated UI frequency, UI volume, cost of self-management, and avoidance activities following an educational intervention designed to increases nurses' identification and management of UI in women. Participants also reported they were less bothered by their symptoms. A further uncontrolled trial found reductions in the volume of UI experienced by community-dwelling adults as well as increases in their knowledge and ability to control and cope with their UI after receiving home-based advice from nurse continence advisors (NCAs) who completed a 3-month self-directed education program combined with 75 hours of supervised clinical practice and independent practice.28 The NCAs also subjectively rated reported that 42% (73/174) of patients were moderately improved or continent.
No studies were identified that evaluated the effects of education on patient outcomes in acute care units. However, 2 trials provide evidence of the effects of nursing education about UI on patient outcomes in rehabilitation units,29,32 both enrolled stroke patients. Frasure32 found no significant differences in the pre- and postintervention frequency of UI among 29 stroke patients in a neuroscience ward, despite a reported 50% improvement in the implementation of a prompted voiding protocol. Similarly, Thomas and associates29 reported no difference in the presence or absence of UI at 6- and 12 weeks poststroke based on the ICIQ-UI Short Form. Overall, 161 (39.7%) of participants were continent at discharge; 72 (44%) in group 1; 51 (41%) in group 2; and 38 (31%) in group 3. We believe this trial provides the strongest evidence of the effects of educating nurses on patient outcomes in a rehabilitation setting.
DISCUSSION
We completed a systematic review of the efficacy of UI education on nurses' and NAs' knowledge, attitude toward, practices, and patient outcomes and found mixed evidence. Whilst nonrandomized and before-after studies tended to report positive effects, these findings were not confirmed in RCTs. Moreover, in 2 large RCTs, education combined with facilitation29 and/or audit and feedback26 did not produce statistically significant differences in patient-related outcomes. This finding is inconsistent with research in other areas about facilitation, audit, and feedback.35 There are many possible reasons for the lack of between-group statistical improvements; they include (1) variability in the methodological quality of the trials, (2) variability in the strength and quality of the educational intervention, (3) the inherent limitations of education in achieving behavioral and organizational change, and (4) limitations of current outcome measures.
Variability in Methodologic Quality
Trial quality is an important consideration. Most researchers described a complex educational intervention comprising multiple components, rendering it difficult to identify the relative merit of one component over another. The methodological quality of the studies varied from 25% to 75% for uncontrolled trials and 25% to 100% for controlled trials. Reasons for the low quality ratings included small sample sizes, multifaceted and complex interventions, reliance on subjective reporting measures, low response rates, high dropout rates, incomplete outcome data, potential bias in recruitment methods, nonvalidated data collection instruments, and incomplete reporting. Future research on the topic should be informed by reporting guidelines, such as the Consolidated Standards for Reporting Trials (CONSORT) statement.36
Variability in Educational Interventions
The strength and quality of the various educational interventions is also an important consideration. Educational content should align with evidence-based recommendations for the management of UI. We assert that these education programs should also be theoretically informed and accommodate the specific context. For example, we assert that RN and NA education should differ. It is unclear if this was the case in most of the trials that included NAs. Nursing assistants are the first-line managers of UI and other bladder and bowel disorders in most nursing homes37 and are key to the uptake of interventions to optimize continence.20 Further research is required to elucidate the differential education RNs and nurse assistants require to assess and manage UI.
Limitations of Education in Achieving Behavioral and Organizational Change
While improved knowledge and attitudes are important enablers of change, continuing education alone is unlikely to produce sustained changes in practice or corporate culture.38 Similarly, evidence suggests that increased knowledge about UI does not necessarily translate into improvements in practices that improve patient outcomes.39,40 RNs and NAs are not solely responsible for the quality of continence care. Few studies accounted for or addressed contextual factors or facilitation/support of practice change in the setting.
Based on review findings, a key barrier to the uptake of educational recommendations about UI could be a lack of audit and feedback data about patients' objective continence status. Thomas and colleagues29 found staff members' belief in the efficacy of the intervention was a critical factor in the uptake of a systemic voiding program. However, nurses are not always able to identify patients' actual continence status,23 possibly due to the challenge of obtaining objective data. Further attention should be given to improving methods to increase nurses' ability to accurately identify patients' frequency and severity of UI.
Limitations of Current Outcome Measures
Outcomes used in the studies included in this systematic review were primarily based on the implicit goal of cure or reducing rates of UI and/or FI, which may be unrealistic for some people. A large proportion of individuals, and particularly those with chronic degenerative neurological conditions and/or a poor prognosis, are not amenable to restoration of continence.41 Ostaszkiewicz42 argues education programs for RNs and NAs should place equal value on helping people adjust to changes in bodily function that affect their identity, autonomy, control, and independence. Stated simply, we assert that researchers should evaluate measures of care as well as cure.
CONCLUSION
All RNs and NAs should be equipped with the depth and breadth of knowledge needed to prevent and actively manage UI. Further well-designed trials are required to determine the specific effects of education on nurses' and NAs' continence care practices and patients' continence outcomes.
KEY POINTS
* Education improves nurses' and NAs' knowledge about UI; however, the most effective forms of education that affect practice and patient outcomes are not known.
* The lack of statistically significant changes in practices and patient-related outcomes observed in controlled trials may be attributable to variability in the methodological quality of the trials, the strength and quality of the educational intervention, a reliance on education to achieve behavioral and organizational change, and/or the selection of outcome measures.
* Future trials on the topic should be informed by contemporary reporting guidelines.
ACKNOWLEDGMENTS
We kindly thank Ms Rachel Bush for her contribution to the research and Professor Trisha Dunning for their helpful comments on the article. We also acknowledge funding support from the Continence Nurses Society Australia (Victorian/Tasmanian Branch) for the 2017 Research/Project Scholarship. The research was conducted through the Centre for Quality and Patient Safety Research at Deakin University.
REFERENCES