Development of central venous access device-associated skin impairment assessment instrument.

AIM: To develop an assessment tool to assess the severity and healing of skin impairment with the central venous access device.
DESIGN: Delphi technique.
METHODS: The instrument domain list was developed through a systematic literature review and semi-structured interviews. Experts from China evaluated the relevance and significance of these items in assessing the degree of skin impairment surrounding central venous access device sites through two Delphi rounds. The APA Style JARS checklist for this article was used.
RESULTS: For the systematic literature review, 28 articles were included to develop the wound assessment instrument. From the articles and interview contents, 15 criteria were selected based on reporting frequency. After further screening via in-depth discussion, the central venous access devices associated with the skin impairment assessment tool were refined to include 14 major domains. Through a two-phase Delphi process, 71 items in 12 domains were ultimately retained.

INTRODUCTION

Central venous access devices (CVADs) are indispensable in the daily clinical management of many patients. CVADs are routinely placed to provide patients with a wider range of treatment options, such as long‐term infusion of anticancer drugs for target treatment; convenient blood collection, long periods of parenteral nutrition management; the delivery of vasopressors or sufficient fluids, or the monitoring of central venous pressure for critically ill patients (Gorski et al., 2021). It is estimated that approximately 8% of hospitalized patients require CVADs during hospitalization, and more than 5 million central venous punctures are performed per year in the United States (Duwadi et al., 2019). Although catheter placement into the central vein is a common procedure performed in many clinical settings, severe insertion and retention complications associated with CVADs, such as bloodstream infections, skin lesions, phlebitis and thrombosis, are still prevalent (Jiang et al., 2020; Mimoz et al., 2015; Pu et al., 2020; Ullman et al., 2019).

BACKGROUND

A multistage research project, organized by the World Congress of Vascular Access (WoCoVA), bridged the obvious gap in the definition of CVAD‐associated skin impairment (CASI) in 2017 (Broadhurst et al., 2017). CASI was defined based on the definition of medical adhesive‐related skin injury (MARSI). This definition was also written into the ‘Infusion Therapy Standards of Practice’ in 2021 (Gorski et al., 2021). Skin complications associated with the use of CVADs have been reported to be a frequent occurrence. The prevalence rates of skin impairment as a result of moisture‐related dermatitis, local infections, contact dermatitis and MARSI associated with CVAD sites, such as redness, rash, vesicle, itchiness or bruising, ranged from 11.7%–36%, and in China, ranged from 3.23%–19.7% across reported studies (Chan et al., 2017; Ullman et al., 2019; Wang et al., 2019; Zhao et al., 2018). Moreover, the presence of CASI often results in substantial adverse impacts, such as patients' discomfort, device failure, increased risk of infections and considerable healthcare costs (Ullman et al., 2015; Wall et al., 2014).

The use of assessment tools is associated with improvements in CASI management. Standardized management of CASI consists of systematic evaluation, appropriate (or individualized) prevention and long‐term follow‐up, with every aspect critical to the overall success (LeBlanc & Baranoski, 2011; McNichol et al., 2013). The systematic evaluation of skin impairment mainly depends on the professional clinical knowledge and the practical skills of healthcare researchers and medical workers, including complete history‐taking, thorough examination and exact clinical judgement (Broadhurst et al., 2017). Assessment tools are important to consider for conducting clinical evaluations accurately and completely. Nevertheless, the existing tools used to evaluate CASI are inaccurate and incomplete, according to the literature. Timsit and colleagues (Timsit et al., 2009) reported data results from the randomized controlled trials. Nurses used the International Contact Dermatitis Research Group Scale to describe the skin condition of the catheter site. Although the scoring system consisted of four levels, it can only assess skin redness, swelling, blisters and a spreading reaction (Timsit et al., 2009). This scoring system was also used in subsequent skin condition assessments by other investigators (Buetti et al., 2020; Günther et al., 2016; Mimoz et al., 2015; Timsit et al., 2012), but there were some limitations, such as less detailed and ordinal response format to report the size of blisters. In other studies, unvalidated local institutional grading was used to assess skin status (Curtis et al., 2015; Su et al., 2017). No targeted instruments have been developed to assess the condition of skin impairment at CVAD sites.

Therefore, to improve the quality and safety of CVADs use, we systematically developed a wound assessment tool to evaluate skin conditions. The goal of this instrument is that members of the health care team can easily apply the instrument, which contains only the most relevant information, and evaluate the severity and healing of skin impairment during catheter management reliably and accurately.

METHODS

The APA Style JARS checklist (Levitt et al., 2018) was followed in reporting on this study (Appendix S1). The following procedures were used to develop a new CASI assessment tool. The content of the instrument was developed through a systematic literature review, which was conducted to inform potential domains that could be used to describe CASI. After identifying domains from the literature review, semi‐structured interviews were conducted with 10 nurses with experience in vascular access care to ensure that their descriptions were reflected in the assessment instrument. Feedback provided by the nurses was integrated into the instrument to produce the archetypal assessment instrument. Experts from China evaluated the relevance and significance of the items to assess the degree of skin impairment surrounding a CVAD site through two Delphi rounds. A multi‐stage, qualitative‐method research approach was utilized to produce the content of the evaluation instrument and ensure its appropriateness for clinical settings.

Literature searching

A systematic literature search was performed by three reviewers in March 2019. The databases PubMed, Scopus, Web of Science, Embase, Cochrane Central Register of Controlled Trials, Cochrane reviews, Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Joanna Briggs Institute Library were searched through database‐specific strategies. After the search, the results were imported into a document management software, and the duplicate records were automatically deleted. Two authors independently reviewed the titles and abstracts to retain records that might be relevant. All abstracts selected by the reviewers were then subjected to a full manuscript review. All articles that met the inclusion criteria were included in the systematic review. Disagreements were resolved through discussion with a third researcher. The eligibility criteria for the study included case reports, expert opinions, randomized controlled trials (RCTs), observational studies that described the objective symptoms and signs of CASI. Studies were excluded if the method of assessment was not described or if they were exclusively conducted with children and/or adolescents under 18. Articles about the clinical manifestations and assessment of phlebitis, drug extravasation and dermatoses, such as pyoderma and varicella, were excluded from this review. Additionally, this review also excluded research on non‐human subjects and articles not available in English. The domains of the wound assessment tool were extracted from the eligible studies identified in the systematic reviews based on reporting frequency of clinical symptoms and signs of skin injury.

Semi‐structured interviews

A representative sample of clinical nurses with experience in vascular access care was invited to participate in semi‐structured, in‐depth, one on one interviews based on a descriptive qualitative research approach (Holt & Hughes, 2021). The aim of the interviews was to determine the most relevant and important categories that characterize CASI in a clinical practice setting. Before the interviews, we formulated several major questions that were continually revised during the interviews. Any interview excerpt describing CASI and discomfort was marked as a domain. Each interview was conducted in a quiet, private room and recorded. The transcription was performed after each interview, and was reviewed and coded by a researcher under the support of qualitative data analysis software (NVivo 11). In order to collect data accurately, themes were continuously revised to facilitate the emergence of new content until saturation of data was reached.

Establishment of the item pool

The drafted indicators were generated according to clinical studies of CASI, reviews, case reports, expert consensus and the nurses' interview results. These items were integrated into a preliminary instrument to assess key characteristic of skin lesions and develop appropriate rating content. Through detailed and in‐depth discussion among the members of experts, the initial instrument was examined for reasonableness, simplicity and redundancy to accommodate the Chinese vascular access care setting, and to improve clinical usability and efficiency.

Delphi technique

The Delphi technique is considered to be one of the most commonly used research approaches to guide the improvement of content validity of an assessment instrument, which demonstrates its suitability for this study. The Delphi method was used to collect quantitative and qualitative data from an advisory group during two consecutive rounds of questionnaires (McPherson et al., 2018). Purposive sampling methods were used to form an advisory group of multidisciplinary experts. The expert panel recruitment criteria were as follows: (a) they had a Bachelor's degree or above, (b) they had at least 10 years of experience in skin/wound or catheter management and (c) they were willing to participate in the Delphi procedure. To avoid the homogeneity of the expert panel, the panellists were purposively recruited from different district polyclinics and medical universities. A professional and secure survey website was used that sent automated reminders to panellists to complete the questionnaire over a period of 1 week.

The content of the questionnaire included three parts. The first part consisted of the complete instrument with detailed domains and response options. In each round, the experts were instructed to rate each statement according to the degree to which they considered it to be relevant for the assessment of CASI via a 5‐point Likert scale (1 = strongly irrelevant, 2 = irrelevant, 3 = neutral, 4 = relevant and 5 = strongly relevant), and were encouraged to offer additional recommendations in blanks. The second part collected demographic information of the panellists (i.e. age, occupation, work seniority, education level and professional title). The third part was a self‐assessment of judgement criteria (Ca) and familiarity (Cs) (Li et al., 2018; Wang & Si, 2011).

Statistical analysis

The databases were created in Microsoft Office 365 Excel, and the statistical analyses were carried out in IBM SPSS, version 25.0. All comparisons were two‐sided, and p values ≤.05 were considered statistically significant.

Ethical considerations

The study was conducted at the comprehensive Grade 3A hospital in southeastern China, and ethical approval was obtained from the Medical Research Ethical Committee. The research conformed to the Declaration of Helsinki, and all participants voluntarily participated and provided informed consent.

RESULTS

Article selection

The results of literature review are displayed in the PRISMA flow chart (Figure 1). A total of 4,637 articles were originally identified during the searches. After removing the duplicates, the reviewers screened 3,988 titles and 137 abstracts. Subsequently, 64 articles were selected for full‐text review, of which 39 were excluded based on the inclusion and exclusion criteria. Three additional full texts were identified by reviewing the references of articles and were added to the analysis. In total, 28 articles were included in the systematic review for the development of the wound assessment tool. These publications included 11 RCTs, 6 observational studies, 4 case reports, 3 expert opinion articles, 3 expert consensus documents and 1 quasi‐experiment. All studies described the specific manifestations of skin injury. Ten studies described wound assessment instruments, with 4 utilizing the International Contact Dermatitis Research Group system to assess the severity of contact dermatitis; 3 using MARSI classification criteria; 1 using a grading scale for skin tears and 2 using their local institutional assessment scales to assess skin injury. In most studies, only a few aspects of skin injury were described, which is not all‐inclusive. In addition, the terminology of assessing CVAD‐associated skin impairment varied widely among studies, and no unified definition could be confirmed.

FIGURE 1

Flow diagram of literature search and selection process

Nurse interview

Ten nurses with vascular access care experience participated in semi‐structured face‐to‐face interviews. After interviewing 8 nurses, data saturation was reached. All participants were women with a bachelor's degree, and 80% had more than 10 years of work experience. The demographic characteristics of all participants are summarized in Table 1.

TABLE 1

Demographics and clinical characteristics of nurses interviewed

Description	N(%)
Age (mean ± SD) (y)	38.5 ± 8.54
Sex (female)	10 (100)
Education level (bachelor)	10 (100)
Professional title
Senior nurse	4 (40)
Nurse‐in‐charge	3 (30)
Co‐Chief nurse	2 (20)
Chief nurse	1 (10)
Specialty nurse (Vascular access)
Yes	4 (40)
No	6 (60)
Work seniority (mean ± SD) (y)	17.3 ± 10.36
<10 years	2 (20)
10–20 years	4 (40)
≥20 years	4 (40)

Item pool

From the articles and interview contents analysed, 15 criteria were selected based on the reporting frequency. After further screening via in‐depth discussion with the panel of experts, the CASI‐assessment tool was refined to 14 major domains that were considered as the most critical factors in monitoring and evaluating CASI healing. The refined instrument included domains for wound size, position and shape, skin colour, depth, size of vesicle and erythema, swelling and hardness of the catheter exit site, warmth, type/amount of exudate and pruritus and pain degree. Each of these domains had a detailed descriptive subscale. The initial instrument was sent to the advisory group panellists via email.

Advisory group

Thirteen experts from general hospitals and universities were recruited to complete the questionnaire and then asked to recommend questionnaire to colleagues or friends who could participate in the study through a sampling technique called ‘snowball sampling’. Using participants' personal social network for snowball sampling when contacting difficult‐to‐reach target groups can shorten the recruitment time and create a sufficient size and heterogeneous sample (Wohl et al., 2017). After this process, the final advisory group consisted of 19 scholars, clinical nurses and doctors from the provinces of Zhejiang, Guangzhou, Shanghai, Chongqing and Yunnan, who had expertise in wound management, vascular access, geriatric care, critical care and oncology care (Table 2).

TABLE 2

Characteristics of the Delphi panel experts

Description	Round 1, N (%)	Round 2, N (%)
Age (mean ± SD) (y)	45.63 ± 5.34	46 ± 5.24
Professional title
Professor	4 (21.05)	4 (22.22)
Associate professor	13 (68.42)	12 (66.67)
Nurse‐in‐charge	2 (10.53)	2 (11.11)
Occupation
Teacher	1 (5.27)	1 (5.56)
Nurse	14 (73.68)	13 (72.22)
Doctor	4 (21.05)	4 (22.22)
Education level
Doctor	4 (21.05)	4 (22.22)
Master	4 (21.05)	3 (16.67)
Bachelor	11 (57.90)	11 (61.11)
Work seniority (mean ± SD) (y)	24.95 ± 7.02	25.44 ± 6.87
<20 years	6 (31.58)	5 (27.78)
20−30 years	7 (36.84)	7 (38.89)
≥30 years	6 (31.58)	6 (33.33)

Initiative coefficient

The expert panel response rate was used to represent the initiative coefficient. The response rate of the questionnaire was 100% (19/19) in round one, and 94.74% (18/19) in round two. According to the literature, a 70% or higher response rate is regarded very good (Li et al., 2018). In the first round, seven panellists presented constructive suggestions for revisions on domains and items. The term ‘pink’ was removed from the colour domain. Panellists recommended changing the content of the ‘amount of exudate’ domain, as the type of dressing may vary from gauze to transparent dressing. Additionally, the panellists suggested that domains like ‘wound size’ should be described in greater detail. The modifications of the initial instrument were adjusted based on the feedback from experts, and then the second round Delphi was conducted.

Authority coefficient

The authority coefficient (Cr) was used to evaluate the degree of authority of experts' opinions. Cr usually depends on three factors: the academic level of experts, the judgement criteria (Ca) and familiarity (Cs). Cr = (Ca +Cs)/2 (Wang & Si, 2011). Ca was the sum of clinical experience, theoretical analysis, literature references and intuition. Familiarity (Cs) consisted of 5 levels, that is very familiar, familiar, neutral, unfamiliar and very unfamiliar, which corresponded to values of 1.0, 0.8, 0.6, 0.4 and 0.2, respectively. In general, Cr≥0.7 is considered to indicate acceptable reliability. Our results showed that the Cr values were 0.879 and 0.886, respectively.

Coefficient of concordance

Kendall's coefficient of concordance (Kendall's W) was used to quantify the strength of consensus among the multiple raters of the CASI assessment tool. The value ranges from 0 (no agreement) to 1 (perfect agreement) (Melinosky et al., 2021; Stanhope et al., 2021). In our study, the two rounds of Delphi results showed that Kendall's W values were 0.394 and 0.402 for the domains and 0.279 and 0.330 for the items. All the results were statistically significant (Table 3).

TABLE 3

Kendall's coefficient of concordance

Round	Part	N	Kendall's W	Chi‐square	DF	p
Round 1	Dimensions	19	0.394	97.357	13	.000*
	Items	19	0.279	270.440	51	.000*
	Total	19	0.286	353.258	65	.000*
Round 2	Dimensions	18	0.402	86.761	12	.000*
	Items	18	0.330	444.905	75	.000*
	Total	18	0.318	503.246	88	.000*

Abbreviation: DF, degrees of freedom.

p ≤ .001

Level of concentration of expert opinions

The level of concentration of expert opinions is calculated by the arithmetic average (Mj) and full mark ratio (Kj) (Wang & Si, 2011). Mj is the average score of each item based on the evaluations of all experts. Kj is the ratio of the number of experts who evaluate the indicator to be very relevant to the total number of experts. The evaluation indexes are screened by the boundary value method, and Mj, Kj and the coefficient of variation (Vj) are calculated according to the importance (or relevance) of each item (Cui et al., 2018). The coefficient of variation (Vj) refers to the degree of coordination of experts' evaluation for indicators. The exclusion criteria of the indicators are: Kj and Mj < mean (MD) ‐ standard deviation (SD) and Vj > MD + SD (Hu et al., 2019). To prevent important indicators from being eliminated, the indicators that did not meet the requirements of the three functions were included. For some unqualified or controversial items, the research group discussed them and determined whether they would be retained in accordance with the principles of comprehensiveness, scientificity and feasibility (Table 4) (Wang & Si, 2011). The first round of Delphi results showed that the Kj, Mj and Vj cut‐off values were 43.66, 3.84 and 0.29 for the domains and 35.89, 3.44 and 0.41 for the items, respectively. In the second round, the cut‐off values were 50.65, 4.01 and 0.27 for the domains and 45.22, 4.10 and 0.23 for the items, respectively (Table 5).

TABLE 4

Mj, Kj, Vj, SD and I‐CVI of the dimensions and items

Results Delphi round 1 (N = 19)							Results Delphi round 2 (N = 18)
Abbreviated item content	Kj	Mj	SD	Vj	I‐CVI	Result	Abbreviated item content	Kj	Mj	SD	Vj	I‐CVI	Result
A Anatomical position of skin impairment	73.68	4.68	0.58	0.12	0.95	Stay	A Anatomical position of skin impairment	83.33	4.83	0.38	0.08	1.00	Stay
Marking relevent position on the picture.	57.89	4.32	0.89	0.21	0.74	Stay	Marking relevent position on the picture.	16.67	3.28	1.13	0.34	0.39	Remove
B Wound shape	26.32	3.00	1.45	0.48	0.32	Remove	Left neck	94.44	4.94	0.24	0.05	1.00	Stay
Irregular	21.05	2.79	1.47	0.53	0.26	Remove	Right neck	94.44	4.94	0.24	0.05	1.00	Stay
Slender	10.53	2.58	1.30	0.51	0.21	Remove	Left chest	94.44	4.94	0.24	0.05	1.00	Stay
Round	10.53	2.26	1.24	0.55	0.16	Remove	Right chest	94.44	4.94	0.24	0.05	1.00	Stay
Square	10.53	2.21	1.28	0.58	0.16	Remove	Left upper limb	88.89	4.89	0.32	0.07	1.00	Stay
C Size	84.21	4.84	0.37	0.08	1.00	Stay	Right upper limb	88.89	4.89	0.32	0.07	1.00	Stay
Intact	42.11	3.74	1.45	0.39	0.63	Stay	Left groin	83.33	4.83	0.38	0.08	1.00	Stay
<25 cm2	52.63	3.84	1.46	0.38	0.63	Change	Right groin	83.33	4.83	0.38	0.08	1.00	Stay
25 cm2<S ≤ 100 cm2	31.58	3.58	1.35	0.38	0.58	Change	B Size	83.33	4.83	0.38	0.08	1.00	Stay
>100 cm2	36.84	3.53	1.43	0.41	0.53	Change	Intact	83.33	4.83	0.38	0.08	1.00	Stay
D Size of vesicle	84.21	4.84	0.37	0.08	1.00	Stay	≤1 cm²	83.33	4.83	0.38	0.08	1.00	Stay
No	79.95	4.42	1.30	0.30	0.84	Change	1 cm²<S≤4 cm²	83.33	4.78	0.55	0.11	0.94	Stay
Yes, but limited to the dressing area.	52.63	4.16	1.07	0.26	0.74	Change	4 cm²<S≤9 cm²	83.33	4.78	0.55	0.11	0.94	Stay
Yes, but exceeded the dressing area.	52.63	4.05	1.27	0.31	0.74	Change	9 cm²<S≤16 cm²	77.78	4.72	0.57	0.12	0.94	Stay
E Size of erythra	78.95	4.68	0.67	0.14	0.89	Stay	16 cm²<S≤25 cm²	72.22	4.67	0.59	0.13	0.94	Stay
No	57.89	4.05	1.35	0.33	0.68	Change	25 cm²<S≤36 cm²	72.22	4.67	0.59	0.13	0.94	Stay
Yes, but within the dressing area.	57.89	4.10	1.20	0.29	0.68	Change	36 cm²<S≤49 cm²	55.56	4.44	0.70	0.16	0.89	Stay
Yes, but beyond the dressing area.	52.63	3.95	1.31	0.33	0.63	Change	49 cm²<S≤64 cm²	44.44	4.28	0.75	0.18	0.83	Stay
F Depth	78.95	4.79	0.42	0.09	1.00	Stay	64 cm²<S≤81 cm²	44.44	4.28	0.75	0.18	0.83	Stay
Intact, or with pigmentation	73.68	4.42	1.26	0.29	0.89	Stay	81 cm²<S≤100 cm²	50.00	4.22	1.06	0.25	0.83	Stay
Intact, but with constant redness	78.95	4.68	0.75	0.16	0.95	Stay	S>100 cm²	50.00	4.22	1.06	0.25	0.83	Stay
Only epidermis missing	84.21	4.79	0.54	0.11	0.95	Stay	C Size of vesicle	83.33	4.83	0.38	0.08	1.00	Stay
Both the epidermis and dermis are missing	84.21	4.74	0.73	0.15	0.95	Stay	Intact	83.33	4.83	0.38	0.08	1.00	Stay
Damage or necrosis of the tissues under the dermis	89.47	4.74	0.93	0.20	0.95	Stay	≤1 cm²	83.33	4.83	0.38	0.08	1.00	Stay
G Colour	73.68	4.63	0.68	0.15	0.89	Stay	1 cm²<S≤4 cm²	66.67	4.56	0.70	0.15	0.89	Stay
Normal	73.68	4.26	1.48	0.35	0.84	Stay	4 cm²<S≤9 cm²	66.67	4.56	0.70	0.15	0.89	Stay
Pink	26.32	3.05	1.64	0.54	0.53	Remove	>16 cm²	66.67	4.56	0.70	0.15	0.89	Stay
Red	78.95	4.53	1.07	0.24	0.84	Stay	D Size of erythra	77.78	4.72	0.57	0.12	0.94	Stay
Purple	78.95	4.47	1.26	0.28	0.89	Stay	Intact	83.33	4.83	0.38	0.08	1.00	Stay
Black	57.89	4.00	1.45	0.36	0.73	Stay	≤1 cm²	83.33	4.83	0.38	0.08	1.00	Stay
H Swelling	84.21	4.79	0.54	0.11	0.95	Stay	1 cm²<S≤4 cm²	83.33	4.78	0.55	0.11	0.94	Stay
None	52.63	4.16	1.26	0.30	0.84	Stay	4 cm²<S≤9 cm²	83.33	4.78	0.55	0.11	0.94	Stay
Within 2 cm of catheter exit site	63.16	4.37	1.07	0.24	0.84	Stay	9 cm²<S≤16 cm²	77.78	4.72	0.57	0.12	0.94	Stay
Beyond 2 cm of catheter exit site	47.37	4.16	1.12	0.27	0.84	Stay	16 cm²<S≤25 cm²	72.22	4.67	0.59	0.13	0.94	Stay
I Hardness	57.89	4.47	0.70	0.16	0.89	Stay	25 cm²<S≤36 cm²	72.22	4.67	0.59	0.13	0.94	Stay
None	47.37	4.05	1.31	0.32	0.84	Stay	36 cm²<S≤49 cm²	55.56	4.44	0.70	0.16	0.89	Stay
Within 2 cm of catheter exit site	63.16	4.42	0.96	0.22	0.89	Stay	49 cm²<S≤64 cm²	44.44	4.28	0.75	0.18	0.83	Stay
Beyond 2 cm of catheter exit site	57.89	4.42	0.84	0.19	0.89	Stay	64 cm²<S≤81 cm²	44.44	4.28	0.75	0.18	0.83	Stay
J Exudate type	68.42	4.53	0.77	0.17	0.84	Stay	81 cm²<S≤100 cm²	50.00	4.22	1.06	0.25	0.83	Stay
None/dry wound	78.95	4.42	1.30	0.30	0.84	Stay	S>100 cm²	50.00	4.22	1.06	0.25	0.83	Stay
Bloody	73.68	4.53	0.90	0.20	0.84	Stay	E Depth	77.78	4.78	0.43	0.09	1.00	Stay
Serous	73.68	4.58	0.84	0.18	0.89	Stay	Intact, or with pigmentation	83.33	4.83	0.38	0.08	1.00	Stay
Purulent	89.47	4.68	1.00	0.21	0.89	Stay	Intact, but with constant redness	83.33	4.83	0.38	0.08	1.00	Stay
K Exudate amount	63.16	4.37	0.96	0.22	0.79	Stay	Only epidermis missing	88.89	4.89	0.32	0.07	1.00	Stay
None/dry wound	52.63	4.16	1.26	0.30	0.84	Stay	Both the epidermis and dermis are missing	83.33	4.83	0.38	0.08	1.00	Stay
<25% dressing saturation	52.63	4.00	1.33	0.33	0.74	Change	Damage or necrosis of the tissues under the dermis	55.56	4.44	0.70	0.16	0.89	Stay
25%–50% dressing saturation	52.63	4.05	1.27	0.31	0.74	Change	F Colour	83.33	4.83	0.38	0.08	1.00	Stay
51%–75% dressing saturation	47.37	4.00	1.29	0.32	0.79	Change	Normal	83.33	4.83	0.38	0.08	1.00	Stay
76%–100% dressing saturation	57.89	4.11	1.33	0.32	0.79	Change	Red	83.33	4.83	0.38	0.08	1.00	Stay
L Warmth	15.79	3.37	1.01	0.30	0.42	Remove	White/Gray	50.00	4.28	1.02	0.24	0.89	Stay
No	10.53	3.00	1.29	0.43	0.37	Remove	Purple/Dark red	50.00	4.28	1.02	0.24	0.89	Stay
Yes	10.53	3.00	1.29	0.43	0.37	Remove	Black	16.67	3.39	1.24	0.37	0.57	Remove
M Pruritus	57.89	4.16	1.17	0.29	0.68	Stay	G Swelling	55.56	4.44	0.70	0.16	0.89	Stay
None	68.42	4.26	1.33	0.31	0.79	Stay	None	50.00	4.39	0.70	0.16	0.89	Stay
Pruritus without the need to scratch	78.95	4.53	1.02	0.23	0.84	Stay	Within 2 cm of catheter exit site	50.00	4.39	0.70	0.16	0.89	Stay
Pruritus with the need to scratch but without excoriations	73.68	4.53	0.90	0.20	0.84	Stay	Beyond 2 cm of catheter exit site	50.00	4.33	0.84	0.19	0.89	Stay
Pruritus unrelieved by scratching but without excoriations	68.42	4.53	0.84	0.19	0.89	Stay	H Hardness	55.56	4.44	0.70	0.16	0.89	Stay
Pruritus accompanied by excoriations	68.42	4.47	0.90	0.20	0.84	Stay	None	50.00	4.33	0.84	0.19	0.89	Stay
Totally restless	68.42	4.21	1.40	0.33	0.79	Stay	Within 2 cm of catheter exit site	50.00	4.28	1.02	0.24	0.89	Stay
NPain	57.89	4.42	0.77	0.17	0.84	Stay	Beyond 2 cm of catheter exit site	50.00	4.28	1.02	0.24	0.89	Stay
None	73.68	4.42	1.26	0.29	0.89	Stay	I Moisture type	61.11	4.50	0.71	0.16	0.89	Stay
Slight	68.42	4.53	0.84	0.19	0.89	Stay	None/dry wound	50.00	4.39	0.70	0.16	0.89	Stay
Moderate	68.42	4.47	1.02	0.23	0.89	Stay	Perspiration	33.33	3.33	1.46	0.44	0.39	Remove
Heavy	73.68	4.58	0.96	0.21	0.95	Stay	Bloody	94.44	4.94	0.24	0.05	1.00	Stay
							Serous	50.00	4.39	0.70	0.16	0.89	Stay
							Purulent	94.44	4.94	0.24	0.05	1.00	Stay
							J Moisture amount (Frequency of dressing change)	66.67	4.61	0.61	0.13	0.94	Stay
							None/dry wound	83.33	4.78	0.55	0.11	0.94	Stay
							Less than once a day	50.00	4.33	0.84	0.19	0.89	Stay
							Once a day	50.00	4.39	0.70	0.16	0.89	Stay
							More than twice a day	50.00	4.39	0.70	0.16	0.89	Stay
							K Odour	22.22	2.89	1.49	0.52	0.33	Remove
							No	16.67	3.17	1.10	0.35	0.28	Remove
							Yes	11.11	3.11	1.08	0.35	0.33	Remove
							L Pruritus	66.67	4.61	0.61	0.13	0.94	Stay
							None	94.44	4.94	0.24	0.05	1.00	Stay
							Pruritus without the need to scratch	77.78	4.78	0.43	0.09	1.00	Stay
							Pruritus with the need to scratch but without excoriations	66.67	4.67	0.49	0.10	1.00	Stay
							Pruritus unrelieved by scratching but without excoriations	50.00	4.28	1.02	0.24	0.89	Stay
							Pruritus accompanied by excoriations	50.00	4.28	1.02	0.24	0.89	Stay
							Totally restless	72.22	4.67	0.59	0.13	0.94	Stay
							M Pain	66.68	4.61	0.61	0.13	0.94	Stay
							None	88.89	4.89	0.32	0.07	1.00	Stay
							Slight	77.78	4.78	0.43	0.09	1.00	Stay
							Moderate	50.00	4.39	0.70	0.16	0.89	Stay
							Heavy	61.11	4.44	0.98	0.22	0.94	Stay
							Severe	66.68	4.44	1.04	0.23	0.89	Stay

TABLE 5

Exclusion criteria of dimension and item

	Round 1			Round 2
	Mean	Standard deviation	Cut‐off	Mean	Standard deviation	Cut‐off
Domains
Full mark rate (%)	64.66	21.00	43.66	67.95	17.30	50.65
Arithmetic average	4.40	0.56	3.84	4.53	0.52	4.01
Coefficient of variation	0.18	0.11	0.29	0.15	0.12	0.27
Items
Full mark ratio (%)	57.59	21.70	35.89	66.08	20.86	45.22
Arithmetic average	4.07	0.63	3.44	4.52	0.42	4.10
Coefficient of variation	0.30	0.11	0.41	0.15	0.08	0.23

Content validity

Content validity refers to the degree of consistency among content measured by the evaluation instrument and the actual domains. The content validity index (CVI) is the most widely used indicator in the quantitative assessment of content validity (Almanasreh et al., 2019). The CVI is divided into two categories: the item‐level CVI (I‐CVI) and scale‐ level CVI (S‐CVI). The I‐CVI was calculated based on the proportion of items rated as 4 or 5 by the experts on a 5‐point scale (Li et al., 2018). When there are more than 5 experts, the I‐CVI standard should not be <0.78, which indicates that the indicators in the evaluation instrument represent a good degree of consistency regarding the specific areas of concern. Depending on the calculation method, the S‐CVI can be divided into two functions: (a) S‐CVI/UA (universal agreement), that is, the percentage of items that all experts rate as 4 or 5 points; and (b) average S‐CVI (S‐CVI/Ave). As the number of experts increases, the possibility for disagreement increases, so the S‐CVI/UA is sensitive to the number of experts and is therefore not recommended. Although there are three methods to calculate S‐CVI/Ave, the use of the average I‐CVI of all items in the scale is recommended. Researchers have suggested that the S‐CVI/UA should be more than 0.8 and that the S‐CVI/Ave should reach 0.90, indicating that the content validity is good (Almanasreh et al., 2019). After the second Delphi round, the I‐CVI values of the items that we retained all exceeded 0.78 (Table 4). In this instance, S‐CVI/Ave =0.90, which is the criterion for acceptability.

DISCUSSION

In clinical settings, health workers are increasingly aware that CVADs are not only a simple tool to provide therapies for vulnerable patients (Broadhurst et al., 2016). Successful insertion and management of CVADs play an important role in improving service quality and ensuring medical safety. During long‐term CVAD placement, the skin integrity surrounding catheter sites is a substantial challenge, as the skin is repeatedly in contact with disinfectants, dressings, etc. (Mimoz et al., 2015; Ullman et al., 2019). In 2017, a study provided practical references for clinical researchers and health workers to identify, prevent and treat CVAD‐related skin lesions (Broadhurst et al., 2017). However, our literature review found that there was no specialized, targeted tool that had been systematically developed for CASI assessment. In view of the clinical problem of a lack of targeted CASI tools, we began to develop an instrument to fill the existing gap. Healthcare researchers and medical workers could accurately characterize the changes in skin conditions surrounding CVAD sites and assess the severity and healing of the impaired skin quickly via this instrument, which is an adequate and consistent measurement tool, has cultural compatibility and is accessible and affordable.

The Delphi technique is an effective and structured method to refine and collect opinion and expertise on the assessment of CASI from a range of experts (McPherson et al., 2018). In this study, a group of experts that constituted the Delphi panel reached consensus on the content of the CASI assessment tool. After the two Delphi rounds, the response rates exceeded 70%, indicating that the Delphi participants had positive attitudes towards this research (Li et al., 2018). The authoritative coefficients were >0.70 in the two Delphi rounds. Because the expert authority coefficient is related to familiarity and the judgement criterion, our results showed that expert authority was high and reliable (Wang & Si, 2011). Since the representativeness of an expert panel is based on the experts' ability and the quality of their expertise rather than the number of experts, to ensure the reliability of expert opinions, the Delphi participants who had an abundance of theoretical knowledge and clinical practice experiences in long‐term work were invited to participate. They were familiar with the research topics and had professional recognition in CVAD management or wound healing. The increase in the Kendall's W value indicated the increase in the expert consensus on the CASI assessment tool, and the indicator system was accepted by panellists (Wang & Si, 2011). The excellent content validity indicated that the CASI instrument contained the most relevant content that could be used to identify the severity of skin impairment (Yusoff, 2019). Our research considered the important symptoms and clinical manifestations of CASI, ensuring that the topic was comprehensively analysed.

Clinician might express different insights on wound care due to the variation in their clinical experience and expertise (Erbay Dalli & Kelebek Girgin, 2021). To avoid gaps in inspecting skin conditions surrounding the CVAD site, researchers and clinicians should consider using standardized, validated assessment instruments to examine skin integrity regularly (Broadhurst et al., 2017; Greatrex‐White & Moxey, 2015). The contents of our assessment instrument provide clinicians with evidence‐ and consensus‐based guidance to assess the clinical manifestations of CASI qualitatively and quantitatively, including the extent of erythema, swelling, exudation, itching and blisters. Our findings indicated that the panellists' recommendations for the item descriptions allowed the instrument to describe the relevant symptoms in detail. Overall, the instrument is acceptable and valid and includes an ordinal response format to describe symptom severity and healing for CASI.

In the future, the tool will be applied in CASI clinical assessment of Chinese patients, including inpatients and outpatients. The departments initially planned to carry out an evaluation activity include gastrointestinal surgery, respiratory medicine, breast surgery and PICC clinic. After further validation, it will continue to be extended to the whole hospital. Differences in assessment results can be compared between nurses with different work seniority and between nurses and doctors. Due to the low number of ethnic groups of other skin colours in the clinical setting in China, inter‐ethnic comparisons are not yet possible.

LIMITATIONS

There were several study limitations. First, most of the members of the expert panel were from the same province, which might have affected the generalizability and heterogeneity of the results. Second, some subjectivity was required when analysing some skin impairment domains, so this tool is more suitable for professional personnel. Third, we have yet to validate the tool in clinical practice. Further investigation is required to assess the validity and reliability.

CONCLUSION

The CASI assessment instrument is an important, targeted tool to guide clinicians in identifying and diagnosing the severity and healing of skin injury around CVAD sites. More prospective clinical studies are needed to verify the feasibility and stability of the tool.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS

LM and YYQ contributed to the study design, the data collection, data interpretation and writing of the manuscript. ZP contributed to the study design and writing of the manuscript. KZW and GXY contributed to the statistical analysis, data interpretation. ZCH contributed to writing of the manuscript. All authors contributed to the preparation of the manuscript and approved the final submitted version.

App S1

Click here for additional data file.