2.1 Ostomy skin tool 2.0

Current UK guidelines recommend that healthcare professionals assess the severity of PSCs using the Ostomy Skin Tool (OST).^{24, 25} PSCs are assessed based on the area affected and severity of symptoms for three different categories: discolouration, erosion, and tissue overgrowth (DET).²⁴ The overall severity of the PSC is then quantified based on the composite DET score, defined on a scale from 0 to 15, and subjects are subsequently assigned to one of the four severity categories: ‘normal skin’ (DET = 0), ‘mild’ (DET = 2–3), ‘moderate’ (DET = 4–6), or ‘severe’ (DET = 7–15).²⁶

One limitation, however, of the OST is that it is reliant on the presence of discolouration, to some degree, in the peristomal area. In the absence of any discolouration, the DET score will be zero, and the peristomal skin will be categorised as ‘normal skin’.²⁶ However, 75% of participants in a recent multinational survey, for people with an ostomy, reported that they still experienced other symptoms of PSC despite the absence of discolouration in the peristomal area.²⁷

In light of the above, a modified OST, known as the OST 2.0, has recently undergone validation; Figure 1 provides further detail on the processes of the decision tree. Using a decision tree, PSCs are categorised based on predefined symptoms including peristomal ulcers; peristomal bleeding; peristomal weeping; pain, itch, or burning sensation; and discolouration.²⁸ The OST 2.0 assesses a greater range of symptoms than the OST and uses an artificial intelligence-based approach to assess the area of peristomal skin affected by discoloration. As a result, in addition to the existing categories of ‘mild,’ ‘moderate,’ and ‘severe’ PSCs (but excluding ‘normal skin’), the OST 2.0 includes an additional ‘no treatment required’ category. Individuals presenting with exclusively peristomal discolouration will be categorised as having ‘no treatment required’ if they have ≤10 cm² discoloured skin or ‘mild’ if they have >10 cm².²⁸

2.2 Survey development

This study involved two surveys designed to assess the HCRU associated with PSCs; a survey assessing the expert opinion of nurses with experience working with stoma patients in the UK, and a patient survey assessing the personal experience of UK patients with an ileostomy. Healthcare professionals (HCPs) and patients are expected to have complementary knowledge regarding the HCRU associated with PSCs. Patients may be able to provide a more accurate description of their day-to-day use of ostomy appliances and supporting products, while HCPs may be able to provide a clearer overview of the treatment that patients typically receive for different PSC symptoms.

A steering committee was formed, which consisted of four clinical experts (a dermatologist and three SCNs); the patient survey was additionally reviewed by two stoma patients. Materials relating to the surveys were validated by the steering committee, to ensure that the questions were appropriate and understandable.

2.3 Patient survey

Patients from the UK were identified through the study sponsor's research database; these individuals had consented to be contacted with respect to future market research relating to stoma care as part of the Coloplast Online Research Engine and for data to be published on an aggregated level. The survey was distributed by a third party and data were anonymised before being shared with the study sponsor. To qualify for inclusion in the study, survey respondents were required to be 18 years or older with an established ileostomy (at least six months post-surgery). A total of 105 patients completed the survey, all of whom met the inclusion criteria.

The survey questionnaire covered patient characteristics including type of appliance used, peristomal body profile, history of PSCs, and number of PSCs experienced in the last three months, as well as asking patients about their use of ostomy appliances and supporting products when not experiencing a PSC. Patients were also asked to recall up to three previous PSC instances that they had experienced within the past three months. Three months was selected based on feedback from the steering committee, as it was agreed that patients could be expected to accurately remember their HCRU over this length of time. In relation to these PSC instances, patients were asked about the symptoms experienced, the duration of symptoms, the use of ostomy appliances, use of supporting products, nurse consultations, and referrals to other specialists. Each PSC was classified as either ‘no treatment required, “mild,” “moderate,” or “severe”’ as per OST 2.0 criteria.²⁸ The exact wording in the OST 2.0 was adapted to match the setting of the survey, asking patients about PSC symptoms in the last 3 months (as opposed to PSC symptoms because they last changed their appliance).

2.4 HCP survey

HCPs were identified from a database administered by the study sponsor, and the link to the online survey was distributed directly via email. To qualify for inclusion in this study, survey respondents were required to be nurses working in stoma care, with experience working with ileostomy patients in the UK within the last year. All HCPs with whom the survey was shared were qualified specialist SCNs, registered with the Nursing and Midwifery Council, and working in the UK. HCPs were reimbursed for their time spent completing the survey at fair market value.

In the first part of the questionnaire, HCPs were asked to estimate patients' use of ostomy appliances at baseline (while not experiencing a PSC) and while experiencing any of the PSC symptoms measured as part of the OST 2.0. In the second part of the questionnaire, HCPs were asked to estimate patients' HCRU while experiencing different sets of symptoms that aligned with each category of the OST 2.0 decision tree.²⁸ The questions covered the duration of symptoms, number of consultations with a nurse, probability of stoma revision, proportion of patients referred to specialists, and proportion of patients prescribed PSC-related medications.

A total of 51 HCPs responded to the survey; one respondent did not pass the screening questions. Of the 50 respondents, 22 completed only the first part of the questionnaire, while the remaining 28 completed the full questionnaire.

2.5 Ethical considerations

Ethical approval for this study was not sought as the study was conducted as market research to understand, but not change, service delivery for ileostomy patients. Furthermore, patients were identified independently of the National Health Service (NHS), and HCPs were identified by virtue of their professional roles; both were noted as exclusions by the NHS Health Research Authority, precluding the need for NHS Research Ethics Committee Approval (http://www.hra-decisiontools.org.uk/ethics/).²⁹ As for all studies involving human participants, basic ethical principles were followed, including gaining informed consent and respecting patient confidentiality.

2.6 Data analysis

Data were cleaned to exclude respondents who did not pass the screening questions and any missing answers (including when the ‘Don't know’ response was selected). Summary statistics were generated for each question. In the patient survey, subgroup analyses were also conducted, including appliance type (one-piece or two-piece), peristomal body profile (inward, outward, or regular), and PSC history (within the past 3 months, more than 3 months ago, or never). Results for patients who reported never experiencing any PSC symptoms were assigned to a distinct ‘no PSC’ category.

A student's t-test was used to compare the frequency of changing across different appliance types (one-piece appliance, two-piece pouch, and two-piece baseplate). In the patient survey, the frequency of changing appliances was compared between baseline and while experiencing PSC symptoms using a student's t-test. In the HCP survey, the same comparison was performed using a Welch's t-test to account for differences in variance as more respondents completed the first part of the survey (n = 50) than the second part of the survey (n = 28). In both surveys, the proportion of patients using each type of supporting product was also compared between baseline and while experiencing PSC symptoms using a Chi-squared test.

To provide inputs for HCRU by PSC severity for the cost analysis, data by symptom endpoint of the OST 2.0 were transformed into results for each severity category using a weighted average.²⁸ Weightings were based on the distribution of endpoints within each severity category from data collected as part of the Coloplast Ostomy Life Survey 2019.²⁸ For symptom endpoints for which no data were available, the weighted average was calculated across the remaining endpoints within the severity category.

2.7 Cost analysis

An estimated cost per PSC instance was calculated for each PSC severity category based on the reported HCRU relative to baseline (eg, cost of additional appliances used while experiencing a PSC in comparison to when not experiencing a PSC). Costs for ostomy appliances were based on a basket of comparators from the most common ostomy product manufacturers in the UK (Coloplast, Hollister, Salts, Convatec & Dansac), as published in the prescription cost analysis.³⁰ The cost of nurse consultations and referrals was identified from the NHS Reference Costs and the PSSRU Unit Costs of Health and Social Care, and medication costs were sourced from the BNF and the drug and pharmaceutical market information tool.^31-38 Supporting product costs were based on products available from Coloplast, as published in the prescription cost analysis.³⁰ All products were selected from a single brand as a simplification, however given the low variation and overall costs of supporting products this is not expected to materially impact the results. The patient survey identified the proportion of patients using supporting products; however, the survey did not include a question relating to the number of supporting products used per week. This input was therefore sourced from Meisner et al., a population-based cost modelling study based on the OST.²⁰

To assess which parameters the PSC cost estimates are most sensitive to, a deterministic sensitivity analysis was also conducted in which HCRU inputs were varied by ± 20% of their mean value.

3.1 Health care resource use

3.1.1 Patient survey

The patient survey included 105 respondents reporting on a total of 172 PSCs. The majority of respondents used a one-piece appliance (74%), and they commonly had an outward (43%), or regular (41%) body profile. Most respondents had experienced a PSC in the past three months (76%), with almost half having experienced three or more in this time period (46%). Only 11% of respondents reported having never had a PSC. No variation in ostomy appliance use was found across patients with different body profiles or PSC histories. Respondents reported more frequent changing of appliances for two-piece pouches than for one-piece appliances, whereas the changing frequency for two-piece baseplates was significantly lower than for one-piece appliances (P = 0.006; Table 1).

TABLE 1. Key results from the patient and HCP HCRU surveys.

OST 2.0 category		One-piece		Two-piece			All appliance types
		n	Appliances per week	n	Pouches per week	Baseplates per week	n	Duration of symptoms (weeks)	n	Visits to an SCN	n	Online/telephone consultations with a nurse	n	Home care nurse visits
Patient survey	No PSC	72	4.59 (3.83)	26	5.71 (3.26)	2.43 (1.27)	NA	NA	NA	NA	NA	NA	NA	NA
	No treatment required	0	–	0	–	–	0	–	0	–	0	–	0	–
	Mild PSC	–	5.40	–	–	5.00	–	0.82	–	0	–	0	–	0
	Moderate PSC	–	9.68	–	–	4.70	–	0.54	–	0	–	0	–	0
	Severe PSC	–	10.82	–	–	6.39	–	2.10	–	0.20	–	0.06	–	0.03
HCP survey	No PSC – No history of PSCs	49	4.50 (1.70)	48	5.58 (2.26)	2.54 (0.60)	NA	NA	NA	NA	NA	NA	NA	NA
	No PSC – History of PSCs	49	5.43 (3.96)	47	6.34 (3.87)	3.57 (3.89)	0	–	0	–	0	–	0	–
	Mild PSC	28	6.96 (2.41)	28	7.13 (2.23)	5.62 (1.91)	–	1.67	–	1.70	–	1.40	–	0.81
	Moderate PSC						–	2.63	–	2.44	–	1.90	–	1.53
	Severe PSC						–	3.61	–	3.06	–	2.27	–	1.94

• Note: Results are provided as weighted averages across the decision tree categories when presented by PSC severity category, and as mean (SD) for all other results. Visits to an SCN are reported as a proportion of the week. No n-values are available for results calculated as weighted averages. In the HCP survey, respondents were asked to estimate appliance use for patients experiencing any PSC symptoms, rather than by severity category. For the purposes of calculating costs, two-piece pouch use was based on the higher value out of baseline two-piece pouch use and PSC two-piece baseplate use from the patient survey.
• Abbreviations: HCP, healthcare professional; NA, not applicable; OST, ostomy skin tool; PSC, peristomal skin complications; SCN, stoma care nurse.

Of the PSC instances reported by patients, 81% were severe (n = 139), 13% were moderate (n = 22) and 6% were mild (n = 11); no patients reported symptoms categorised as no treatment required. For both one-piece appliances and two-piece baseplates, the number of appliances used per week was higher when experiencing PSC symptoms than when not experiencing PSC symptoms (both P < .001), particularly for the most severe symptoms (Table S1). The proportion of respondents using supporting products at baseline varied substantially across product categories (Figure 2), and there was no clear trend in the proportion of patients using supporting products when experiencing PSCs compared with baseline or across PSC severity categories. On average, only patients with the most severe symptoms of peristomal ulcers, sores, or peristomal bleeding reported that they consulted with a nurse as a result of their symptoms (Table 1).

3.2 Cost of PSCs

The patient survey was used to inform estimates of ostomy appliance use and supporting product use for the cost calculations, while the HCP survey was used to inform estimates of symptom duration and treatment received by PSCs, including nurse consultations, referrals, medication, and stoma revision. The HCP survey asked respondents to estimate the number of nurse consultations that patients would receive across three different categories: visits to a nurse, home care nurse visits, and non-face-to-face (online or telephone) consultations. There is some uncertainty regarding whether respondents may have considered these categories separately or in combination. Taking a conservative approach, the base case cost analysis counts only in-person visits to an SCN, which was the most frequent type of nurse consultation reported by HCPs (Table 1).

On the basis of these estimates, a total HCRU cost was calculated for each PSC severity. Total HCRU is estimated to cost £258, £383, and £505 for the duration of a mild, moderate, or severe PSC, respectively (Figure 3), as a result of both direct treatment costs and increases in ostomy appliance use compared to when not experiencing a PSC.

The impact of the nurse visits assumption was explored in a scenario analysis. When all nurse consultation types were included, the estimated cost per PSC increased by approximately 36% relative to the base case. Alternatively, if all nurse visits reported by the HCPs were assumed to be non-face-to-face visits, reflecting the potential situation during the COVID-19 pandemic, PSC costs reduce by approximately 63%. However, it is expected that this reflects an extreme rather than likely scenario, and would very likely result in an increase in other forms of HCRU, a longer duration of PSC symptoms, or a reduction in the quality of care experienced by patients.

Deterministic sensitivity analysis results indicate that nurse visits were the largest cost driver across all three PSC severity categories (Figure 3). The top four parameters across all three categories were the cost of nurse visits, one-piece appliances, revision/re-surgery and GP referral (although the ordering varied across PSC severity categories). All other parameters resulted in a variation in the total PSC cost estimate of less than ± £2.50 (Figure 4).

As a scenario analysis, the PSC costs were recalculated using the number of nurse visits recalled by patients and reported in the patient survey, which were considerably lower compared to the visits reported by HCPs. As a result, the PSC costs for mild, moderate, or severe were estimated to be £51, £91, and £161 respectively.

4.1 Limitations

This study has important strengths, such as consulting both patients and HCPs to gain a clear picture of HCRU and having a large sample size of HCPs used to determine PSC-associated HCRU.^{16, 20} To our knowledge, this study is also the first to investigate HCRU specifically in the UK and to focus on patients with an ileostomy.^{6, 7, 10, 14} However, there are also some important limitations to acknowledge.

The lower response rate in the second part of the HCP questionnaire may have been because of the increased complexity of the questions, despite the steering committee previously agreeing that the questions were appropriate. The response rate may have also been impacted by the COVID-19 pandemic, as HCPs would have less time available for engaging in research activities. However, despite these factors, to our knowledge, 28 SCNs represent the largest sample of HCPs used to determine HCRU associated with PSCs to date.^{16, 20}

Secondly, patients were asked to report up to three previous instances of PSCs that they had experienced within the past three months, meaning that the results are reliant on accurate recall. On the basis of feedback from the steering committee, patients could be expected to accurately remember PSCs and associated HCRU over this length of time, however, it is likely that the results are subject to some recall bias.

A third limitation is that the patient survey did not provide an estimate for the HCRU associated with symptoms categorised as ‘no treatment required’, as no patients reported a PSC instance that met the criteria for this category. Patients with minor PSC symptoms may be less likely to consider these as indicative of a skin disorder, instead considering these symptoms to be a normal consequence of having a stoma.^{5, 7, 10, 12-14} HCPs provided non-zero estimates for the treatment of PSCs categorised under the new OST 2.0 as ‘no treatment required’ (Table 1). This may reflect the fact that the current OST does not include this category, and therefore, under the current assessment, nurses would likely expect all symptoms identified as a PSC to receive treatment. However, according to the OST 2.0, these patients would not be expected to need treatment and would therefore likely be associated with minimal additional costs relative to someone who was not experiencing a PSC.²⁸ The omission of such patients from this cost analysis is therefore unlikely to result in a meaningful underestimation of PSC costs, but rather represents a conservative estimate.

Finally, previous cost analyses have assumed an increased use of supporting products while experiencing a PSC.^{19, 23} This was not included in the present cost analysis as the patient HCRU survey did not ask about the frequency of supporting product use while experiencing a PSC. Therefore, the current analysis may underestimate the costs associated with supporting product use while experiencing a PSC. However, given the small unit costs for supporting products, this is similarly not expected to have a substantial impact on PSC cost estimates.