Cost-Utility Analysis of Negative Pressure Wound Therapy Compared With Traditional Wound Care in the Treatment of Diabetic Foot Ulcers in Iran

Vahid Alipour, PhD; Aziz Rezapour, PhD; Mehdi Ebrahimi, MD; Jalal Arabloo, PhD


Wounds. 2021;33(2):50-56. 

In This Article

Materials and Methods

This economic evaluation study was conducted in 2016. In order to simulate the health status of patients with DFU (NPWT = 100; TWC = 100) under different health conditions, the Markov model approach was used in this study. The Markov model is generally suitable for examining chronic and recurring conditions (similar to the alternatives investigated in this study).[20,21] The initial structure of the Markov model in this study was based on models used in previous studies.[5,16] The initial model also was adjusted for the treatment process of Iranian patients with DFU. Moreover, information on the outcomes and transition rates of different health diagnoses were extracted from the related published sources.[5,16] The time period of this study was considered to be 1 year and based on monthly cycles (ie, 12 cycles). The Markov model structure used in this study is shown in Figure 1. The health diagnoses considered in this study included the following: uninfected, infected, infected post-amputation, healed, healed post-amputation, amputation, and death. The transition between different health conditions was determined based on the DFU development, so that a person with DFU may clinically go through different health conditions, as shown in Figure 1. Based on the natural history of DFU, there are some clinical impossibilities to the transition between some diagnoses (ie, there is no possibility for patients to be directly transferred from an infected state to a healed state or from death to another health state). The analysis of cost-effectiveness and sensitivity for uncertain parameters was performed using TreeAge Pro 2011 (TreeAge LLC). An expert panel validated the model, including clinicians specialized in diabetes, health economics, and health policy.

Figure 1.

Diabetic foot ulcer model structure.

Transition Rates

Monthly transition rates between different health states indicated the probability of moving between 2 possible health conditions during 1 month. For example, with NPWT, the probability of change of an uninfected wound of the diabetic foot into an infected wound is 3.6%, but the probability is 4.3% when using TWC. Also with NPWT, the probability that the infected wound would result in amputation is 1.1%, vs 3.8% with TWC. Correspondingly, these probabilities were derived from studies by Flack et al[5] (1000 hypothetical patients) and Whitehead et al[16] (1000 hypothetical patients) on NPWT and wound dressing strategies (Table 1).

Resource use

The resources used for each health diagnosis were estimated monthly and separately based on inpatient (ie, inpatient days, nursing visits, antibiotics, orthopedic appliances, amputation, prostheses, and materials/dressings) and outpatient care (ie, physician visits and tests). The resources used to estimate the inpatient costs were obtained from medical records of 200 patients who were admitted to university-affiliated hospitals in Tehran, Iran. The resources used based on various health diagnoses are shown in Table 2.


The present study investigated the cost-effectiveness of NPWT compared with TWC treatment strategies from the perspective of health care providers. To calculate monthly costs per each health state for NPWT and TWC strategies, unit costs of inpatient days, nursing visits, physician visits, amputation surgery, and DFU dressing were extracted from the Relative Value of Health Services,[22] which were then multiplied by their relevant monthly resource usage. Costs of antibiotics, orthopedic appliances, prostheses, and tests were obtained directly from the patient's medical records, with the exception of the cost of mortality, which was not considered in this study (Table 3).

Over 1 month, NPWT was performed in 3-day sessions 4 times. During treatment, the amount of negative pressure varied according to the progress of treatment (12 times/month). TWC was implemented every day (30 times/month). To calculate monthly dressing, wound expert, and nursing costs per patient for these 2 treatment strategies, the authors multiplied dressing change per month by cost per dressing change (Table 4).

Afterward, all costs were converted into US dollars using an average exchange rate based on the rate claimed by the Central Bank of Iran (in 2016, US $1 = 32 000 rials)

Utility Weights

Utility weights needed to calculate QALYs were derived from the published literature (Coffey et al,[23] n = 2048 subjects with type 1 and type 2 diabetes; Whitehead et al,[16] n = 1000 hypothetical patients). If there was a variation in utility values in the studies, the mean of utility values was considered as the utility weight. Utility weights in a plausible range in the sensitivity analysis section were considered and then analyzed. Notably, the utility value for mortality was considered zero (Table 5).

Time Period

A 1-year time period was considered for calculation of costs and outcomes; therefore, the discount rate was not considered for payoffs.


Model effects were expressed as QALYs (utility weights of health diagnosis × length of life in that condition). Economic evaluation findings were reported in terms of ICER.

Sensitivity Analysis

In order to analyze sensitivity, uncertainty parameters (ie, costs and effects) were inserted with a sensible range in a Tornado diagram for examining the sensitivity of model results to the variables of costs and effects. Because cost variables may be significantly affected by the economic conditions of Iran and utility values may be differently reported in various studies, costs and effects have changed in the range of ± 30% and ± 20%, respectively.