Is Noncontact Normothermic Wound Therapy Cost Effective for the Treatment of Stages 3 and 4 Pressure Ulcers?
Is Noncontact Normothermic Wound Therapy Cost Effective for the Treatment of Stages 3 and 4 Pressure Ulcers?
Chronic pressure ulcers are significant health problems, especially for patients in long-term care facilities. Pressure ulcers reduce quality of life and may lead to infection, pain, and death. Estimates of the prevalence of pressure ulcers in nursing home patients range from 3 to 28 percent,1–3 with an annual incidence of 2 to 13 percent.4–6 Reducing the proportion of nursing home residents with pressure ulcers is receiving increasing public health attention and is one of the Healthy People 2010 Objectives.7
While prevention of pressure ulcers should be the primary goal, over the past decade, new adjunctive therapies have become available to treat pressure ulcers.8 These therapies include the use of electric energy,9 external negative pressure,10 low air-loss beds,11 various dressings,12 skin substitutes,13 and growth factors.14 Commonly, however, standard of care consists of cleansing the ulcer, using moisture-retentive dressings, providing a pressure-reducing surface, repositioning the patient, and debriding necrotic tissue.15
Another novel treatment is noncontact normothermic wound therapy (NNWT),* which is a semiocclusive, moisture-retentive wound cover that is warmed to 38 degrees C to deliver radiant heat, which maintains wound temperatures closer to normothermia. Increasing the temperature of wound tissues increases skin oxygen tension16–18 and increases oxygen delivery due to increased perfusion.19,20 These conditions encourage tissue healing21–24 and resistance to infection.25 NNWT also reduces the inhibitory effect of chronic wound fluid26 and stimulates cell proliferation in vitro.27 NNWT has been shown to accelerate healing of chronic venous stasis ulcers.28,29
From a Food and Drug Administration (FDA) regulatory perspective, NNWT can claim increased blood flow and increased oxygen at the wound site. As preliminary studies of wound healing efficacy with NNWT are now complete, the cost effectiveness of NNWT compared to standard care can be examined. The goal of this study was to develop a Markov model to estimate the incremental cost effectiveness of NNWT compared to that of the current standard of care from the societal perspective.
Cost-effectiveness analysis produces a ratio. The numerator takes into consideration the additional costs that one intervention imposes over another. The denominator considers the incremental improvement in health-related quality of life calculated as quality-adjusted life years (QALY).
Our base case focused on a 72-year-old continent man living in a nursing home with a two-month-old stage 3 pressure ulcer (no undermining) in the ischial area measuring 4cm x 4cm. This base case was chosen to be representative of a common clinical scenario. The base case represents our best estimate for each clinical variable. We completed a secondary economic analysis of stage 4 pressure ulcers.
Overview. We reviewed the literature to determine the rates of healing and complications associated with treatment of pressure ulcers, computed associated incremental costs, created a Markov model,30 assumed a societal perspective for the analysis as recommended by an expert panel,31 estimated quality of life for relevant health states, and performed a sensitivity analysis to evaluate the impact of changing key variables. A Markov model is built to simulate the progression of hypothetical patients through various health states related to their illness. The computer model provides a tool to deal with multiple clinical uncertainties and can be run multiple times varying parameters to address “what if” questions.
Patients were modeled to be in one of six mutually exclusive health states: stage 3 ulcer, stage 4 ulcer, healing wound, closed wound healed back to normal, complications requiring hospitalization, and death. Our disease model simulated the progression of an institutionalized patient with a stage 3 pressure ulcer treated either with NNWT or standard care, into and out of these different health states. For the stage 3 pressure ulcer model, all patients started in the stage 3 ulcer health state. We developed a second Markov economic analysis of stage 4 pressure ulcers. All patients in the stage 4 model started in the stage 4 ulcer health state.
A simplification of the Markov model structure for stage 3 pressure ulcers is outlined in Figure 1. Patient progression was divided into eight-week time cycles over a period of 40 months. Forty months was chosen because it allows the model to consider that some pressure ulcers persist for more than three years. We also modeled a 20-month time period. A six-month period would have been too short because it would have not included pressure ulcers that did not respond to initial therapy. Choosing a longer time period, such as 80 months, would have been unnecessarily long. An eight-week cycle time was chosen as it most closely approximated the time of a therapeutic treatment. Age-specific death rates were obtained from 1998 life table data from the National Center for Health Statistics (http://www.cdc.gov/nchs/data/nvsr/nvsr48/nvs48_18.pdf).
Treatment options studied. Standard care was assumed to follow guidelines published by the Agency for Healthcare Quality and Research. In the modeling, this current standard of care consisted of cleansing the ulcer, using moisture-retentive dressings, providing a pressure-reducing surface,32 repositioning the patient eight times per day, and debriding necrotic tissue. As recommended by the manufacturer, NNWT therapy consisted of NNWT administered three times a day for one hour at a time with a minimum of one hour in between warming sessions, cleansing the wound, utilizing a pressure-reducing surface, and repositioning the patient eight times per day.
Literature review. We followed published guidelines33,34 in compiling studies on pressure ulcers. We searched for clinical trials of pressure ulcer healing using electronic searches of MEDLINE (English-language journals with studies conducted on human subjects from 1986 to June 2000). “Pressure ulcer” was entered as Medical Subject Heading and as a text word. In addition, a manual search performed by screening citation lists in review articles yielded another 18 articles. We had no limitations on whether the clinical trial was controlled or randomized or on age group. Abstracts presented at national meetings related to chronic wounds in 1998 and 1999 were manually searched. Where abstract data were incomplete, we contacted the primary author to request further information. Finally, we attempted to identify unpublished data by contacting content experts and study authors.
The contents of 187 abstracts or full-text articles that were identified during our literature search were read in full. We included only controlled trials that were at least four weeks in duration and required that each trial include at least one of the outcome measures described by Price.35 These efficacy endpoints include percentage of patients healed, percentage change in wound area, absolute change in wound area, total area healed, or rate of complete wound healing. Data extraction was performed by the lead author. The following study characteristics were recorded: first author’s name, year of publication, and country of origin; mean age, age range, total number, and gender of patients; presence of coexisting disease; design details, including blinding (open, single, or double) and type of control treatment; and study duration.
The incidences of progressing through the health states were estimated from these articles and converted to bimonthly transition probabilities (Table 1 for stage 3 pressure ulcer and Table 2 for stage 4 pressure ulcer). Given the small number of studies on NNWT, sources of progression data to describe the probability of transition were limited. When there were no empirical data to base transition probabilities, we made estimates based on the available data.
Wound healing endpoint. In the Markov model, we used the probability of wounds healing after eight weeks of therapy as the efficacy measure. These probabilities were estimated based on results of studies summarized in Table 5. Since not all studies included this endpoint, we also analyzed absolute percentage change in wound area differences. For example, if the experimental group of patients receiving NNWT healed 50 percent of the wound after eight weeks while the control group healed 20 percent of the wound, we assumed that healing was 2.5 times (50%/20%) faster with NNWT.36 We averaged these healing rates across the three studies that did have this endpoint, weighting them for number of patients enrolled (see sensitivity analyses below for addressing variation in healing rates among studies).
The model we developed reflects that lower-stage (superficial or partial-thickness pressure ulcers) have more rapid improvement than do higher-stage (deep or full-thickness pressure ulcers).37–39 Healing rates for older (greater than one year) wounds are unavailable in the literature. We did not include an operative cure (i.e., surgical reconstruction with muscle or cutaneous flaps) in the model as more consensus is needed to develop criteria for selecting those individuals most likely to benefit from surgical management.40Measurement of costs. We used a “bottom up” cost methodology to estimate direct medical costs for wound care treatment and complications using data gathered from the literature. Cost of treatment included all costs for treating patients as they progressed through the health states until the ulcer healed or the hypothetical patient died (Table 3). Costs related to nursing were computed by including time per dressing, total nursing time per day, and average cost per nursing hour. Other direct costs included supplies and equipment (see Table 3 footnotes for more detail).41 Complication costs were summed (e.g., antibiotic treatments for systemic infections attributable to the pressure ulcer). Physician professional costs were included.
NNWT consists of a warming element to which a disposable wound cover is attached. The warming element is provided at no out-of-pocket cost, while the disposable wound cover costs $240.00 per wound cover to acquire. We assumed the wound cover to be used around the clock, with the warming element applied three times (with a minimum of one hour between applications) a day to heat the wound as recommended by the manufacturer. We assumed the wound cover would be used for two days, prior to needing to be replaced, for a daily cost of $120.00. For an eight-week cycle in the Markov model, the total acquisition cost for the NNWT wound cover equaled $6720.00.
All costs are reported in year 2000 U.S. dollars. We discounted all future costs and QALYs at three-percent per annum in the base case scenario.
Measurement of health-related quality of life. QALYs include a length of time component (e.g., one year) and a quality-of-life component (i.e., utility). Health utility is the numerical valuation of one’s quality of life on a linear scale from 0.00 (death) to 1.00 (perfect health). For example, one QALY for an individual in perfect health (with a utility = 1.0) for one year (QALY = 1) is considered equivalent to two years in a health state with utility = 0.5 (QALY = 1).
We were unable to find primary data in the literature for quality-of-life assessments for patients with pressure ulcers.42,43 For this reason, we instead used the Rosser classification of illness states, or Rosser index, to form quality-of-life weights (utilities) for each health state in our model.44
The quality of life of a patient with a pressure ulcer was determined by assigning levels of disability and distress to each health state. Thus, the Rosser index has two dimensions, disability and distress, and 29 possible health states. The change in health status, or quality adjustment, was combined with the life expectancy of patients to form QALYs. The Rosser and Kind valuation matrix was applied to obtain mean values for patients’ changes in health status.45 The utilities assigned via this method were checked with another scale, the Health Utilities Index-2 multiattribute utility function.46 This scale has dimensions for hearing, speech, vision, emotion, pain, ambulation, dexterity, cognition, and self care. The estimates of utilities used in the model were compiled from these two instruments (Table 4).
Measurement of life-years. Baseline mortality rates were based on published life tables.47,48 These were modified to reflect that failure of a pressure ulcer to heal is associated with a tripling of mortality.49,50Calculation of the cost-effectiveness ratio. Incremental cost effectiveness was calculated by comparing the incremental costs and QALYs with NNWT versus the incremental costs and QALYs obtained with standard treatment. We examined the accuracy of our disease model by comparing the predictions of the model with endpoints observed in clinical studies of treatment of pressure ulcers with those of standard care, not necessarily studies involving NNWT.
Sensitivity analyses. The impact of varying our assumptions over a reasonable range was explored with sensitivity analyses.51 The ranges of estimates for each parameter encompassed the ranges reported in the literature. Because only single studies were available for some parameters, ranges of possible effectiveness were estimated for the purposes of sensitivity analyses. These ranges are outlined in Table 1.52
The probabilistic sensitivity analysis considered uncertainties in all probabilities, utilities, and costs simultaneously. Mean values for the two end points (incremental decrease in costs, incremental increase in QALY) were calculated for results of N = 10,000 Monte-Carlo simulations (@Risk 4.0, Palisade Corporation, Newfield, New York). Triangular distributions were used for parameter values, with the mode being the base case and the 5th and 95th percentiles being the lower and upper limits of the ranges reported in Table 1.
Our modeling predicted the total expected cost of standard-care treatment for a stage 3 pressure ulcer to be $20,874.00.
Randomized, nonblinded, clinical trials (total number of patients receiving NNWT = 72, total number of control patients = 59) have found that noncontact, normothermic wound therapy, when used in combination with repositioning and a pressure-reducing surface, decreased the surface area of the stage 3 and 4 pressure ulcers by an average of 2.5 fold (SD 59%) (Table 5). A prospective, randomized trial found the eight-week healing rates for stage 3 ulcers are 71 percent with NNWT and 54 percent for standard care (Whitney study in Table 5). No complications related to NNWT were reported.
At these healing rates, our modeling for stage 3 pressure ulcers found that for a 40-month time frame NNWT saves $6,630.00 (SE $98.00) and increases QALYs by 0.10 (SE 0.0005) as compared to standard care (Figure 2).
For stage 4 pressure ulcers, NNWT saves $15,216.00 (SE $186.00) and increases QALYs by 0.14 (SE 0.001) (Figure 3). The majority of these effects for both types of pressure ulcers occur in the first 20 months (Table 6). For example, our modeling predicts that after 20 months, 88 percent of patients receiving standard care would be expected to be healed, while 95 percent of patients receiving NNWT would be expected to be healed.
Given the uncertainty of input variables, sensitivity analyses were conducted by creating distributions of the estimates based on the ranges outlined in Table 1. The variables that had the largest impact on the modeling were the total cost per day of treatment with standard care, the probability of healing to a normal closed wound with standard care, the cost of the complication state, and the acquisition cost of the NNWT wound cover. When the cost of the NNWT wound cover was increased above $421.00, the use of NNWT increased overall cost to society.
By running repeated iterations of the model and with random sampling, a distribution of the cost-effectiveness values is achieved. The simulations show that NNWT is likely to reduce costs to society and increase quality of life for at least 75 percent (SE 0.4%) of patients with stage 3 ulcers. NNWT is likely to reduce costs to society for at least 81 percent (SE 0.4%) of patients with stage 4 ulcers.
We examined the accuracy of our disease model by comparing the predictions of the model for six-month healing rates and total costs for standard care. These values were similar to those observed in applicable studies.
Pressure ulcers among the elderly remain a major health issue in the United States. While prevention should be the primary goal, early recognition of pressure injury, prevention of ulcer progression, attention to nutrition, infection control, and interventions to improve mobility and continence are essential to pressure ulcer treatment.53
Initial, unblinded, randomized trials suggested that NNWT, as an adjunctive therapy, accelerates by 2 fold the healing of stages 3 and 4 pressure ulcers (Table 5). Although NNWT has a higher acquisition cost than traditional wound care products, our cost-effectiveness computer simulation results suggest that NNWT both reduces costs to society and improves quality of life compared to standard care under a wide range of conditions for patients with stage 3 or stage 4 ulcers. This is primarily because the anticipated accelerated wound healing may lead to improved health-related quality of life as well as to utilization of fewer medical resources and fewer complications. These cost and outcome benefits for NNWT are more pronounced for stage 4 pressure ulcers.
From a public health perspective, cost-effectiveness analyses are useful because their measure of benefit (QALY) incorporate quantity and quality of life, permitting comparison of all interventions on a uniform scale. NNWT costs less and has increased QALYs as long as the healing rate with NNWT is greater than 1.5 times better than standard care. This finding is unlike most cost-effectiveness analyses of new technologies, which often find that while the new treatment adds benefit, the new treatment also costs more than standard care. In this situation, medical interventions with a cost effectiveness of less than $50,000.00 per QALY are generally considered to represent acceptable value for money, i.e., cost effectiveness.54
Given the uncertainty of input variables, sensitivity analyses revealed that the variables that had the largest impact on the modeling were the daily treatment costs and the probability of healing to a normal closed wound.
We tested the robustness of our modeling by comparing what the computer simulation would predict with published studies documenting the natural history or progression of pressure ulcers. For example:
1. Total expected cost of standard-care treatment for a stage 3 pressure ulcer was predicted to be $20,874.00 by our modeling. Studies that have measured total costs associated with care of a patient with a pressure ulcer have found the per-ulcer cost can be in that range.55,56
2. According to our model, 65 percent of patients with stage 3 ulcers undergoing standard therapy would be expected to be healed at six months. This is greater than the 45-percent healing rate at six months measured among patients with pressure ulcers at VA long-term care facilities.57 Rates of healing reported in that study are derived from administrative data and, as such, are likely to be underestimates of the true healing rate, which may explain why our model had a greater six-month healing rate. The underestimation may arise as patients whose ulcers healed and who then developed new pressure ulcers were considered not healed. Also, many patients had more than one ulcer and were not counted as healed unless all of their pressure ulcers had healed. Third, healing of pressure ulcers in patients that were discharged was not included. That study also found that 31 percent of patients with stage 4 ulcers would be healed at six months. In another study, a national nursing home chain similarly reported a 40-percent healing rate for stage 3 pressure ulcers and a 34-percent healing rate for stage 4.58 Both of these rates are within a few percentage points of what our model predicts.
3. Our model predicts that the mean length of treatment for a stage 3 ulcer receiving standard care, from initial occurrence to complete healing, will equal 153 days. This is consistent with a wound closure rate of 0.1cm2/day as documented by Ferrell59 for a 16cm2 wound (as per our base case). Also, a study of long-term care patients with pressure ulcers found mean treatment length to equal 116 days.60
Advanced wound care products compete with one another and with other medical treatments for reimbursement from federal and commercial payers. Thus, we also completed a second literature search to compare the efficacy of NNWT to topical negative pressure, another advanced wound care product.61 We found one prospective trial with a total of 24 participants that evaluated the effectiveness of topical negative pressure on 36 chronic wounds.62 However, this study did not report statistical analyses, and healing data for the subset of patients who had pressure ulcers was not reported. Additional well-controlled, blinded, “head-to-head” clinical trials, as well as studies of the use of NNWT in nonresearch protocol settings, are necessary.
In our modeling, we assumed that transition probabilities from one health state to another remained constant as time went on (e.g., probability of healing after the first eight-week time period was equal to the probability of healing in the fourth eight-week time period). However, it may be that the efficacy of both standard and NNWT treatments decreases as a pressure ulcer ages. We simulated decreasing efficacy of both treatments as the pressure ulcer aged and found no significant impact on overall results.
Costs, outcomes, and benefits can be analyzed from different points of view—the patient, the provider, the payer, or society as a whole.63 For example, the cost of a medical service to the payer (insurance company) equals the percentage of charges actually paid by the payer. However, the relevant cost to the patient is the out-of-pocket expense (not covered by insurance) plus other costs (e.g., inability to work) incurred due to illness. A different way to assign a cost for nursing services may have impacted the cost model. For example, from the facility’s perspective, nursing care time may be considered a fixed cost, as staff are paid regardless of whether there is one more or one less pressure ulcer to treat. However, we assumed that having a provider take care of a pressure ulcer is an incremental cost to society, as is commonly done in cost-effectiveness studies, to reflect that from society’s point of view, there is a cost for the provider’s time and expertise.
Our computer simulation study has several limitations. First, it is difficult to properly randomize patients with full-thickness pressure ulcers because of the variability in confounding comorbidities, as well as the variability in types, sizes, and locations of pressure ulcers. Also, there may be variability among facilities and among types of patients in how well standard care is actually delivered. Second, studies of NNWT have small sample sizes but do show more rapid reduction in wound surface area. This may not translate directly to a higher probability of complete wound healing as modeled in this study. Although most studies report changes in surface area as the healing outcome for full-thickness ulcers, reliable measurement of changes in depth or volume is also necessary. Third, malpractice costs (median [range] monetary settlement of $279,000.00 [$25,000.00–$65,000,000.00]), related to the development of pressure ulcers among nursing home residents were not included because of their unknown incidence.64
Also, we were unable to find studies that directly surveyed patients with pressure ulcers to quantify the utility or quality of life associated with having a pressure ulcer. A study of patients with chronic venous leg ulcers did report improvements in pain and other physical health dimensions after treatment.65 However, health-related quality-of-life measures have not been studied in nursing home patients with pressure ulcers, in part because assessments may be difficult to obtain in patients with cognitive impairments.66 Also, many patients with pressure ulcers are treated in the community. Our analysis focused on institutionalized patients. Costs and benefits for patients treated in their homes may be different than the results we present.
For this reason, we approximated utilities by mapping those health states with commonly used multi-attribute scales. Our sensitivity analyses suggest that our approach did not affect our results in that the uncertainty regarding the quality of life associated with pressure ulcers was not an important predictor of the cost effectiveness of NNWT. Further studies are necessary to document the impact of NNWT on improving health-related quality of life and reducing infection rates, reducing pain, and reducing need for mechanical debridement.
In conclusion, noncontact, normothermic wound therapy in patients with stage 3 or 4 pressure ulcers is an economically attractive intervention. With a healing rate 2-fold greater with NNWT than with standard care, under a wide range of conditions, NNWT both improves quality of life and decreases costs to society. Additional, well-controlled, clinical trials as well as studies of the use of NNWT in nonresearch protocol settings are necessary to further define the role of NNWT in the care of chronic wounds.
* Warm-Up® Therapy (Augustine Medical Inc., Eden Prairie, Minnesota)