Clostridial collagenase ointment and medicinal honey utilization for pressure ulcers in US hospitals

Abstract

Aims: To describe the utilization of clostridial collagenase ointment (CCO) and medicinal honey debridement methods in real-world inpatient and outpatient hospital settings among pressure ulcer (PU) patients and compare the frequency of healthcare re-encounters between CCO- and medicinal honey-treated patients.

Materials and methods: De-identified hospital discharge records for patients receiving CCO or medicinal honey methods of debridement and having an ICD-9 code for PU were extracted from the US Premier Healthcare Database. Multivariable analysis was used to compare the frequency of inpatient and outpatient revisits up to 6 months after an index encounter for CCO- vs medicinal honey-treated PUs.

Results: The study identified 48,267 inpatients and 2,599 outpatients with PUs treated with CCO or medicinal honeys. Among study inpatients, n = 44,725 (93%) were treated with CCO, and n = 3,542 (7%) with medicinal honeys. CCO and medicinal honeys accounted for 1,826 (70%) and 773 (30%), respectively, of study outpatients. In adjusted models, those treated with CCO had lower odds for inpatient readmissions (OR = 0.86, 95% CI = 0.80–0.94) after inpatient index visits, and outpatient re-encounters both after inpatient (OR = 0.73, 95% CI = 0.67–0.79) and outpatient (OR = 0.78, 95% CI = 0.64–0.95) index visits in 6 months of follow-up.

Limitations: The study was observational in nature, and did not adjust for reasons why patients were hospitalized initially, or why they returned to the facility. Although the study adjusted for differences in a variety of demographic, clinical, and hospital characteristics between the treatments, we are not able to rule out selection bias.

Conclusion: Patients with CCO-treated PUs returned to inpatient and outpatient hospital settings less often compared with medicinal honey-treated PUs. These results from real-world administrative data help to gain a better understanding of the clinical characteristics of patients with PUs treated with these two debridement methods and the economic implications of debridement choice in the acute care setting.

Keywords:

• Clostridial collagenase ointment
• debridement
• healthcare utilization
• medicinal honey
• pressure ulcer
• re-admission
• Santyl
• wound

Introduction

Pressure ulcers (PUs) represent a considerable societal burden, affecting 2.5 million patients annually in the US and accounting for up to $11 billion in healthcare costs¹. Prevention, management, and treatment for PUs present a number of challenges for hospitals because patients with PUs are often very ill, and the resulting complications, such as pain and infections, as well as longer length of stay² and ongoing follow-up care, account for high healthcare utilization and economic consequences. The high cost burden of pressure ulcers takes on increased importance considering the evolution of healthcare financing from pay for volume to pay for value^3–5.

An important aspect of the treatment for PUs is debridement, which removes dead tissue and any foreign material from the wound, helping to promote healing by reducing the number of microbes, toxins, and other substances^6–8. The methods for wound debridement include enzymatic, autolytic, sharp, mechanical, and biological⁹. The choice of debridement method depends on many factors, including but not limited to the characteristics of the patient, wound, risk of infection, and cost¹⁰. For some wounds, more than one method of debridement may be appropriate⁹.

Enzymatic debridement has clinical advantages over other debriding methods by selectively removing necrotic tissue, while at the same time enhancing tissue proliferation^11,12. SANTYL Clostridial Collagenase Ointment (CCO) (Smith & Nephew Inc., Fort Worth, TX) is the only enzymatic debridement agent currently approved by the FDA, and is indicated for clinical use on dermal ulcers¹³. SANTYL CCO contains a mixture of collagenases and non-specific proteases produced by Clostridium histolyticum fermentation that promotes wound epithelialization by increasing the proliferation of vascular endothelial cells and keratinocytes¹³. CCO contains two strains of collagenase: Collagenase G and Collagenase H, that together debride denatured collagen at seven sites¹⁴. Additionally, since CCO cleaves collagen at multiple sites, it facilitates the production of multiple bioactive peptides, unlike endogenous human collagenases involved in the autolytic process of debridement¹⁵.

Autolytic debridement is a broad category of debridement that can include any agent that uses the patient’s own wound fluids to loosen and liquify necrotic tissue¹⁶. Medicinal honeys are promoted by manufacturers as one such alternative, topically applied, autolytic method of debridement^17,18. However, medicinal honey is not approved as a debriding agent, rather it was originally indicated and approved by the FDA as an antimicrobial therapy/treatment option. The mechanism behind the debriding properties of medicinal honey is its support for the body’s own natural process by which endogenous proteolytic enzymes break down necrotic tissue, cleaving at one site of the denatured collagen strand¹⁴. Additionally, the acidic and high sugar osmotic properties of medicinal honeys may promote autolytic debriding by increasing the release of oxygen from hemoglobin, reducing protease activity, and stimulating lymph outflow^17,19,20. Of note is that autolytic debridement such as medicinal honey is dependent on the patient’s immune system, and, therefore, immuno-compromised conditions may negatively impact a patient’s ability to generate autolysis^21,22.

While there is evidence supporting the debriding properties for both CCO and medicinal honeys, comparative effectiveness data from real-world hospital settings are limited²². Therefore, the objectives of this study were to identify PU patients treated with CCO or medicinal honeys among inpatients and hospital-based outpatients in a sample of US hospitals, to describe the costs associated with these debridement types, and then to compare hospital-based re-encounters for these two debridement methods. The hypothesis was that the clinical properties of CCO described above^11–15 would translate into real-world benefits in terms of lower burden of healthcare resource utilization; specifically, fewer revisits during the follow-up time period compared with medicinal honey.

Methods

Data source

Data for this study were derived from the US Premier Healthcare Database (PHD), which includes de-identified clinical and financial data on more than 6 million annual hospital discharges. The PHD is a complete census of inpatient and submitted outpatient hospital-based visits from over 700 geographically diverse hospitals. The de-identified patient daily service records include demographics, disease states, and billed services for medications, laboratory, diagnostics, and therapeutic supplies. The database also includes the hospital characteristics such as bed size, population served (rural vs urban), geographic location, and teaching status. Patients are tracked with a unique identifier across visits to the same hospital. The database is statistically de-identified and complies with the Health Insurance Portability and Accountability Act of 1996.

Study population

The study population included adult (age ≥18 years) inpatients and outpatients diagnosed with a PU and treated with either CCO or medicinal honey debridement in US hospitals during the time period of January 1, 2012 to September 30, 2015, allowing for up to 180 days of follow-up through March 31, 2016. Patients with PUs were identified using primary or secondary International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9) diagnosis codes for PU (707.0x/707.2x). The index visit was defined as the first qualifying inpatient or outpatient encounter meeting the inclusion criteria. In the re-encounter outcomes analysis, we excluded those with unknown gender, patients with LOS >99^th percentile (95 days), and those who expired during the index visit (total excluded n = 4,812) (see Figure 1).

Figure 1. Exclusions for the re-admission and re-encounter outcomes analysis.

Study variables

The exposures of interest in this analysis were CCO + standard wound care (SWC) and medicinal honey + SWC debridement methods. Patients who received combination debridement methods during the index visit were not included in this analysis. CCO was identified by billing evidence for any of the following in the hospital chargemaster: “SANTYL” or “Enzymatic debridement” or “COLLAGENASE” plus “ointment or topical” or “COLLAGENASE CLOSTRIDIUM” plus “ointment or topical” or (“COLLAGENASE CLOSTRIDIUM HISTOL” and no ICD-9 codes for 728.6 Contracture of palmar fascia or 607.85 Peyronie’s disease). Medicinal honeys were identified using billing evidence for “Medihoney” or “*honey*” (non-nutritional supplement). All billing data in the PHD come directly from the hospital, and represent the treatments recorded during the inpatient or outpatient encounter for billing purposes.

Outcomes included all-cause inpatient readmissions or outpatient revisits to the same hospital at 30, 60, 90, and 180 days post-index discharge date. Visit types recorded during the follow-up period included inpatient hospitalizations, emergency department visits, and outpatient hospital-based visits. Non-surgical debridement costs were calculated as the costs specific to CCO or medicinal honeys, as reported in the hospital billing records. The costs were inflation-adjusted to 2015 USD using the US Department of Labor Consumer Price Index-All Urban Consumers data²⁴.

Covariates included patient demographic information, as well as hospital characteristics, which were recorded from the discharge record at the index visit. Standard demographic, admission type, and discharge status variables were indicated by hospital facilities according to the UB-04 codes. Clinical characteristics such as Charlson Comorbidities²⁵ and risk factors for PU were defined during the index visit, with the exception of prior PU and prior diabetic foot ulcer, which were defined using a 12-month look back period for visits to the same facility. The comorbidities and risk factors for PU were defined using ICD codes and/or billing record searches.

Statistical analysis

Univariate analysis was performed to describe the full PU study population treated with CCO or medicinal honey debridement, non-surgical debridement costs, and to help determine variables to include in multivariable outcome regression models. Data measured on a continuous scale were expressed as mean, standard deviation, min–max, median, and IQR. Categorical data were expressed as the frequency and percentage of patients in the categories. Missing values for categorical variables were recorded as unknown. The top and bottom one percentile of non-surgical debridement costs were Winsorized to account for outlying values. Chi-square analysis was used to compare frequencies of categorical variables, and t-tests were used to compare means for continuous variables. All analysis were stratified by inpatients and outpatients.

Multivariable logistic regression modeling examined the association of type of debridement with re-admission/re-encounter outcomes among patients with PUs. The covariates in the multivariable models were chosen based on prior studies of PU risk factors^26,27, in combination with guidance from stepwise selection (p = .20 for entry and p = .15 to stay in the model). For inpatients, the following covariates were included in the models: age (continuous), Charlson Comorbidity Index (CCI)²⁴, race (white, black, other), sex, admission type (emergency, urgent, elective, other), large hospital (500+ beds vs <500 beds), discharge status (expired, home, hospice, skilled nursing facility (SNF)/long-term care (LTC), transferred, other), urban hospital location, provider geographic area (Midwest, West, South, Northeast), hospital teaching status, All Patient Refined Diagnosis Related Group (APR-DRG) level of severity, APR-DRG risk of mortality, prior PU in past 12 months, malnutrition, diabetes, heart failure, chronic kidney disease, dementia, neoplasm, quad/hemiplegia, PU stage, multiple PUs present, shock POA, vasopressor use, wheelchair use, and LOS (continuous days). For outpatients, the covariates for the models included the following: age, race, gender, admission type, discharge status, urban hospital location, provider geographic area, hospital teaching status, prior PU in past 12 months, sepsis, diabetes, heart failure, chronic kidney disease, neoplasm, quad/hemiplegia, PU stage, and multiple PUs present.

Debridement is typically used on Stages III and IV PUs. However, we found that CCO was used in patients with lower stage ulcers, indicating that possibly there was under-staging of these PUs based on the ICD-9 codes, or that these patients had additional, higher stage PUs (22% of inpatients and 9% of outpatients had multiple PUs recorded). Therefore, patients with Stages I and II ulcers were included in the overall study to preserve the sample, but a sensitivity analysis was conducted that restricted the study population to only those patients with Stage III, Stage IV, or multiple stage PUs recorded. The models used for the sensitivity analysis were identical to those specified above for the overall study population.

Results

The study identified 50,866 patients (48,267 inpatients and 2,599 outpatients) with PUs treated with CCO or medicinal honey debridement at one of 701 US hospitals. Among the inpatients, n = 44,725 (92.7%) were treated with CCO, and n = 3,542 (7.3%) with medicinal honeys (Table 1). CCO and medicinal honeys accounted for 1,826 (70.3%) and 773 (29.7%), respectively, outpatients in the study. The mean age of the inpatient study population was 71.1 (SD = 15.6) years, 49.4% were male, 77.4% had Medicare as the primary payer, 75.6% were admitted on an emergency basis, and 50.1% were discharged to a SNF/rehab/LTC facility. The mean age of the PU outpatient study population was slightly younger (68.9 years, SD = 17.4) and, as expected, less ill overall as indicated by the admission type (53.1% elective), discharge status (69.3% home), and lower prevalence of the comorbidities compared with the inpatients. The mean non-surgical debridement cost was $165 (SD = $210) for inpatients, and $79 (SD = $115) for outpatients.

Table 1. Inpatient and outpatient index visit characteristics by study debridement method—pressure ulcer patients.

	Inpatients							Outpatients
	Total CCO and Medicinal Honey		CCO		Medicinal Honey		p-value	Total CCO and Medicinal Honey		CCO		Medicinal Honey		p-value
	n	%	n	%	n	%		n	%	n	%	n	%
# of Unique Patients	48,267	100.0	44,725	92.7	3,542	7.3		2,599	100.0	1,826	70.3	773	29.7
Age
Mean ± SD	71.1 ± 15.6		71.1 ± 15.63		71.0 ± 15.60		.9534	68.9 ± 17.4		70.0 ± 17.20		66.3 ± 17.61		<.0001
Median	73		73		73			71		73		68
25–75 percentile	61–83		61–83		61–83			57–83		59–84		54–80
Min–Max	18–89		18–89		18–89			18–89		18–89		18–89
Age group
18–49	4,514	9.4	4,174	9.3	340	9.6	.3418	366	14.4	234	12.8	132	17.1	<.0001
50–64	10,423	21.6	9,692	21.7	731	20.6		604	23.2	398	21.8	206	26.6
65+	33,330	69.1	30,859	69.0	2,471	69.8		1,629	62.7	1,194	65.4	435	56.3
Race
White	30,272	62.7	27,585	61.7	2,687	75.9	<.0001	1,846	71.0	1,244	68.1	602	77.9	<.0001
Black	10,905	22.6	10,455	23.4	450	12.7		390	15.0	274	15.0	116	15.0
Other	7,090	14.7	6,685	14.9	405	11.4		363	14.0	308	16.9	55	7.1
Sex
Male	23,827	49.4	22,106	49.4	1,721	48.6	.6053	1,266	48.7	854	46.8	412	53.3	.0023
Female	24,439	50.6	22,618	50.6	1,821	51.4		1,333	51.3	972	53.2	361	46.7
Healthcare coverage
Commercial	4,470	9.3	4,149	9.3	321	9.1	.0034	292	11.2	211	11.6	81	10.5	.3777
Medicare	37,358	77.4	34,546	77.2	2,812	79.4		1,970	75.8	1,391	76.2	579	74.9
Medicaid	4,890	10.1	4,593	10.3	297	8.4		244	9.4	161	8.8	83	10.7
Other Payor	1,549	3.2	1,437	3.2	112	3.2		93	3.6	63	3.5	30	3.9
Admission type
Emergency	36,503	75.6	34,158	76.4	2,345	66.2	<.0001	392	15.1	355	19.4	37	4.8	<.0001
Urgent	6,821	14.1	6,047	13.5	774	21.9		166	6.4	87	4.8	79	10.2
Elective	4,545	9.4	4,181	9.3	364	10.3		1,379	53.1	916	50.2	463	59.9
Other	398	0.8	339	0.8	59	1.7		662	25.5	468	25.6	194	25.1
Discharge status
Expired	4,498	9.3	4,130	9.2	368	10.4	<.0001	3	0.1	2	0.1	1	0.1	<.0001
Home	4,392	9.1	3,995	8.9	397	11.2		1,802	69.3	1,244	68.1	558	72.2
SNF, rehab, ICF, or long-term care	24,186	50.1	22,519	50.4	1,667	47.1		237	9.1	221	12.1	16	2.1
Transferred	2,878	6.0	2,675	6.0	203	5.7		27	1.0	21	1.2	6	0.8
Other	12,313	25.5	11,406	25.5	907	25.6		530	20.4	338	18.5	192	24.8
Comorbidities
Malnutrition	15,291	31.7	14,099	31.5	1,192	33.7	.0087	113	4.4	96	5.3	17	2.2	.0005
Immobility	1,170	2.4	3,772	8.4	244	6.9	.2995	62	2.4	44	2.4	18	2.3	.9015
Quadriplegia Hemiplegia	5,949	12.3	5,555	12.4	394	11.1	.0238	310	11.9	207	11.3	103	13.3	.1528
Neuropathy	4,592	9.5	4,246	9.5	346	9.8	.5914	151	5.8	109	6.0	42	5.4	.5934
Urinary incontinence	1,814	3.8	1,695	3.8	119	3.4	.1951	74	2.9	65	3.6	9	1.2	.0008
Heart Failure	14,975	31.0	13,797	30.8	1,178	33.3	.0028	212	8.2	160	8.8	52	6.7	.0831
Chronic Kidney Disease	16,425	34.0	15,216	34.0	1,209	34.1	.8923	257	9.9	203	11.1	54	7.0	.0013
Dementia	12,033	24.9	11,289	25.2	744	21.0	<.0001	269	10.4	224	12.3	45	5.8	<.0001
Diabetes	22,207	46.0	20,574	46.0	1,633	46.1	.9060	821	31.6	603	33.0	218	28.2	.0157
History of Pressure Ulcer in previous year	8,729	18.1	8,129	18.2	600	16.9	.0658	806	31.0	550	30.1	256	33.1	.1310
Sepsis	17,793	36.9	16,615	37.1	1,178	33.3	<.0001	12	1.2	11	0.6	1	0.1	.1039
Charlson Comorbidity Index - Deyo
0	4,218	8.7	3,925	8.8	293	8.3	.0163	1,066	41.0	709	38.8	357	46.2	.0076
1–2	16,224	33.6	15,073	33.7	1,151	32.5		978	37.6	707	38.7	271	35.1
3–4	14,572	30.2	13,535	30.3	1,037	29.3		399	15.4	290	15.9	109	14.1
5–6	8,389	17.4	7,720	17.3	669	18.9		110	4.2	83	4.5	27	3.5
7+	4,864	10.1	4,472	10.0	392	11.1		46	1.8	37	2.0	9	1.2
Non-surgical debridement cost*
Mean ± SD	$165 ± $210		$175 ± $214		$38 ± $62		<.0001	$79 ± $115		$102 ± $130		$26 ± $29		<.0001
Median	$97	–	$106	–	$23	–		$39	–	$61	–	$17	–
25–75 percentile	$40–$204		$49–$217		$10–$42			$15–$95		$23–$124		$9–$31
Min–Max	$5–$1,274		$5–$1,273		$5–$1,254			$2–$727		$2–$727		$2–$260

* Top and bottom 1 percentile of costs were Winsorized to account for extreme outliers.

PU patients who received CCO were the same mean age as medicinal honey-treated patients in the inpatient setting (71.1 and 71.0, respectively, p = .9534), but were older in the outpatient setting (70.0 vs 66.3 years, p < .0001). CCO-treated inpatients were more frequently admitted on an emergency basis (76.4% and 66.2%, p < .0001) compared to the medicinal honey treatment patients. There were statistically significant differences in the frequencies of some comorbidities, notably dementia was more frequent among the CCO-treated patients in both the inpatient (25.2% vs 21.0%, p < .0001) and outpatient (12.3% vs 5.8%, p < .0001) facilities, and chronic kidney disease in the outpatient setting (11.1% vs 7.0%, 0.0013). Diabetes was equally common among both CCO and medicinal honey-treated inpatients (46.0% and 46.1%, respectively, p = .9060), but more common among the CCO-treated outpatients (33.0% vs 28.2%, p = .0157). The non-surgical debridement cost for inpatients was mean $175 (SD $214) for the CCO-treated patients, and $38 (SD $62) for the medicinal honey patients (p < .0001). For outpatients, the costs were mean $102 (SD $130) and $26 (SD $29), respectively (p < .0001).

Figure 2 shows the distribution of the PU stages for the CCO and medicinal honey treatments among inpatient and outpatient index visits. Overall, 54% of inpatients and 43% of outpatients had stages III or IV ulcers during their index visit (data not shown). CCO patients had higher proportions of Stages III (28% vs 22% inpatients [p < .0001], 24% vs 21% outpatients [p = .0176]) and IV (24% vs 15% inpatients [p < .0001], 16% vs 12% outpatients [p = .0083]) PUs compared with medicinal honey-treated patients. The multivariable models accounted for these differences in PU stage between the patient groups.

Figure 2. Pressure ulcer stages among CCO and medicinal honey-treated patients. * Indicates statistically significant difference (p < 0.05) vs CCO. Ulcers that could not be staged are excluded.

After exclusions were applied for the outcomes analysis, there were a total of 46,054 patients (43,458 inpatients and 2,596 outpatients) who were treated with CCO or medicinal honey and included in the analysis of readmissions and re-encounters. Figures 3 and 4 show the unadjusted proportions of index inpatients and outpatients, respectively, who had inpatient readmissions or outpatient re-encounters through 180-days of follow-up. The medicinal honey patients had a higher frequency of 180-day readmissions, as well as outpatient re-encounters at 30, 60, 90, and 180-day follow-up after inpatient index visits. After outpatient index visits, there were no statistically significant differences in inpatient revisits, but a higher proportion of the medicinal honey patients had outpatient revisits at all follow-up time points.

Figure 3. Inpatient re-admissions or outpatient re-encounters after inpatient visits for CCO and medicinal honey-treated pressure ulcers. * Indicates statistically significant difference (p < 0.05) vs CCO.

Figure 4. Inpatient or outpatient re-encounters after outpatient index visits for CCO and medicinal honey-treated pressure ulcers. * Indicates statistically significant difference (p < 0.05) vs CCO.

Table 2 shows adjusted ORs for re-admissions and outpatient re-encounters for CCO vs medicinal honey-treated PUs after considering a wide range of demographic, clinical, and hospital characteristic differences between the treatments. In adjusted models, those treated with CCO had 14% lower odds for inpatient re-admissions (OR = 0.86, 95% CI = 0.80–0.94, p = .0002) at 180-days of follow-up. There was no evidence for a difference between the treatments in inpatient encounters after outpatient index visits. However, odds for outpatient revisits after inpatient and outpatient index visits were lower for the CCO-treated patients at 30, 60, 90, and 180-days. At 30-days, the odds for an outpatient visit was 23% lower (OR = 0.77, 95% CI = 0.70–0.85, p < .0001) after an inpatient index visit, and 19% lower (OR = 0.81, 95% CI = 0.67–0.97, p = .0223) after an outpatient index visit for the CCO treated patients. After 180-days, the odds were 27% lower (OR = 0.73, 95% CI = 0.67–0.79, p < .0001) for after an inpatient index visit and 22% lower (OR = 0.78, 95% CI = 0.64–0.95, p = .0145) after an outpatient index visit for those treated with CCO compared to medicinal honey.

Table 2. Adjusted* odds ratios for revisits for CCO vs medicinal honey debridement treatments for pressure ulcers.

	Inpatient Index Encounter		Outpatient Index Encounter
	CCO (n = 40,294) vs Medicinal Honey (n = 3,164)		CCO (n = 1,824) vs Medicinal Honey (n = 772)
	OR	95% CI (p-value)	OR	95% CI (p-value)
Inpatient revisits
0–30 days	0.97	0.88–1.08 (.6088)	1.28	0.95–1.72 (.1033)
0–60 days	0.95	0.88–1.03 (.2595)	1.14	0.88–1.45 (.3201)
0–90 days	0.93	0.85–1.00 (.0494)	1.16	0.93–1.47 (.1839)
0–180 days	0.86	0.80–0.94 (.0002)	1.09	0.88–1.33 (.4271)
Outpatient revisits
0–30 days	0.77	0.70–0.85 (<.0001)	0.81	0.67–0.97 (.0223)
0–60 days	0.76	0.70–0.83 (<.0001)	0.79	0.65–0.96 (.0189)
0–90 days	0.76	0.70–0.83 (<.0001)	0.81	0.66–0.98 (.0319)
0–180 days	0.73	0.67–0.79 (<.0001)	0.78	0.64–0.95 (.0145)

* Models for the inpatient index visits included treatment, age (continuous), Charlson Comorbidity Index (CCI)²⁴, race (white, black, other), sex, admission type (emergency, urgent, elective, other), large hospital (500+ beds vs <500 beds), discharge status (expired, home, hospice, skilled nursing facility (SNF)/long-term care (LTC), transferred, other), urban location, provider geographic area (Midwest, west, south, northeast), hospital teaching status, All Patient Refined Diagnosis Related Group (APR-DRG) level of severity, APR-DRG risk of mortality, prior PU in past 12 months, malnutrition, diabetes, heart failure, chronic kidney disease, dementia, neoplasm, quad/hemiplegia, PU stage, multiple PUs present, shock POA, vasopressor use, wheelchair use, and LOS (continuous days). Models for the outpatient index visits included treatment, age, race, gender, admission type, discharge status, urban location, provider geographic area, hospital teaching status, prior PU in past 12 months, sepsis, diabetes, heart failure, chronic kidney disease, neoplasm, quad/hemiplegia, PU stage, and multiple PUs present.

The sensitivity analysis that was restricted to patients with Stages III, IV, or multiple stage PUs included 24,448 CCO and 1,443 Medicinal honey-treated inpatients, and 763 CCO and 258 Medicinal honey-treated outpatients. The results remained in the same direction (CCO patients had lower odds for revisits), and statistically significant for cumulative re-admissions 180-days after inpatient index visits (0.87, 95% CI = 0.78–0.98, p = .0255), and for outpatient revisits at 60- (OR = 0.84, 95% CI = 0.73–0.96, p = .0134), 90- (OR = 0.85, 95% CI = 0.74–0.97, p = .0137), and 180-days (0.81, 95% CI = 0.72–0.92, p = .0013) after inpatient index visits. After outpatient index visits, the results also remained in the same direction (CCO lower odds), but statistically significant only at 60 days (OR = 0.70, 95% CI = 0.50–0.98, p = .0394) for outpatient revisits.

Discussion

This study provided real-world data on the characteristics and outcomes for the inpatient and hospital-based outpatient PU population treated with active enzymatic vs passive autolytic debridement, and compared hospital healthcare utilization between those treated with CCO and medicinal honeys. The differences between the treatments in the clinical and demographic characteristics indicated more severe illness and advanced stage PUs among the CCO- vs medicinal honey-treated patients in the inpatient setting. After accounting for a variety of patient demographic, clinical, and hospital characteristics differences between the treatments, CCO-treated patients exhibited fewer 180-day re-admissions, as well as all-cause re-encounters at all time points compared with medicinal honey-treated patients.

Another retrospective observational study supports the results from the current study. A study that used US Wound Registry data found that CCO-treated patients achieved 100% granulation faster (255 vs 282 days, p < .001) after the index date, and had fewer total visits (9.1 vs 12.6, p < .001) compared with medicinal honey-treated patients²³. These results agree with the current study’s findings of lower odds of re-visits for CCO-treated patients. However, the current study did not include measures of granulation, and cannot attribute the lower odds of revisits to faster epithelialization. To our knowledge, there are no other published studies comparing re-admission outcomes for CCO and medicinal honey methods of debridement.

The main strength of the current study is that it provides real-world information on the PU population treated with CCO and medicinal honey in a large, recent sample of patients from US hospitals. Additionally, it provides new information to healthcare providers balancing the demands of positive outcomes, cost of care, and wound treating roles by setting of care. Debridement in the acute care setting is often an afterthought to more pressing concerns. However, this research found that the debridement-related cost difference between CCO and medicinal honey debridement methods in the acute care setting reflected a relatively small investment in potentially reduced resource utilization in post-acute care (mean CCO and medicinal honey spend accounted for a small 0.5% and 0.1%, respectively, of total inpatient costs overall). Lower risk of re-admissions and hospital outpatient re-encounters for CCO vs medicinal honeys may present an opportunity for hospital providers that participate in Accountable Care Organizations (ACOs), bundled payment, or other risk-based contracts to reduce episode costs.

There are limitations to consider when interpreting results. Importantly, the results are specific to the study sample, will vary by institution, and depend on the re-admission/re-encounter rate for each individual hospital. The ability to identify PUs was limited to the ICD-9 data, which do not provide insight into wound severity, such as wound size, presence of infection, presence of tunneling, etc. Therefore, this analysis does not account for these clinical characteristics of PUs that may also be associated with re-encounters. For patients who had multiple PUs, there is no way to determine if the debridement method was applied to all wounds. It also remains unknown if CCO or medical honey was specifically used for debridement, or for other purposes, and so there may be misclassification bias related to the intended use of CCO or honeys. As this study found that the CCO patients had more severe wounds, it is unlikely that the difference in indication and intended use of CCO vs medicinal honeys would explain the results because the bias would have tended to be more favorable toward medicinal honeys. There might have also been some degree of misclassification in the billing data used to identify the CCO and honey-treated patients. However, one treatment should not be more or less accurately recorded in the hospital chargemaster, and, therefore, should not have biased our results. The completeness of coding for some diagnoses and procedures may be motivated by reimbursement policies or other factors. Therefore, there might be under- or over-reporting of certain comorbidities, PU stage, procedures, and complications. Some of the comorbidities were defined using a look-back period; however, these would only include conditions for patients who consistently return to the same hospital for care. Therefore, there is potential that some pre-existing conditions were missed. Similarly, only revisits to the same hospital were captured, and so re-admissions/re-encounters may be under-estimated in this study population. However, as long as there was not differential under-reporting between the debridement treatments, the overall effect should bias estimates for differences toward the null. Local re-admission practices that differ between facilities that are more or less likely to use each treatment also might have accounted for the differences in this study. The lower odds of re-visits may not be attributed to CCO directly. Specifically, the different reasons why patients receiving each type of treatment were admitted to the hospital were not fully accounted for in the data, and the patients likely differed on unmeasured factors (treatment selection bias), and so confounding is not ruled out as the reason for differences in the outcomes between CCO and medicinal honey. Those who expired during the index visit were excluded from the outcomes analyses, but no information was available on those who expired outside of the hospital. Therefore, the analysis does not account for death as a competing risk in this study. However, the proportion of patients who expired during the index visit was similar for both CCO and medicinal honey patients at the index visit (as shown in Table 1), and so we do not believe that differences in death rates outside the hospital would fully explain our results. Additionally, there was no measure of wound healing available in our data. A study with larger sample size for the medical honey-treated patients would be required for propensity score matching and stratification by reason for hospital admission (e.g. MS-DRG, 3M APR-DRG, etc.) to better control for the treatment selection bias.

Conclusions

Using real-world data from a large sample of US hospitals, the results from this study showed fewer re-admissions and outpatient re-visits for CCO- compared with medicinal honey-treated patients with PUs. These results help to gain a better understanding of both the clinical characteristics of wound patients treated with these two debridement methods and the economic implications of debridement choice in the acute care setting. Patients treated with CCO had more advanced stage wounds recorded, yet had a lower frequency of re-admissions and hospital outpatient re-encounters. As acute care providers and administrators consider the investment and setting of care for the appropriate role for debridement, this research provides new findings on the relative low cost investment and high reward of enzymatic debridement compared with medicinal honey early in a patient’s wound episode.

Transparency

Declaration of funding

This study was funded by Smith & Nephew.

Declaration of financial/other relationships

CW and GD are employees of Smith & Nephew. JD, JG, and RJ are employed by and have equity ownership in Premier, Inc. Peer reviewers on this manuscript have received an honorarium from JME for their review work, but have no other relevant financial relationships to disclose.

Previous presentations

These results were presented in abstract form at the Symposium on Advanced Wound Care (SAWC) Fall meeting, October 20–22, 2017, and at the Academy of Managed Care Pharmacy (AMCP) Nexus 2017, Dallas, TX, October 16–19, 2017.

Acknowledgments

No assistance in the preparation of this article is to be declared.