A comparison of photographic and transparency-based methods for measuring wound surface area.

Objective measurement of wound size is necessary to document change in response to treatment. Such documentation is important both to clinicians concerned with individual patients and to researchers comparing healing responses among groups of patients. The ideal measurement would be accurate, reliable, simple to obtain, economical of time and material, and safe for the patient. Various methods for measuring wound surface area (WSA) of chronic ulcers in patients have been described in the literature. Two of the most commonly used methods involve obtaining a tracing of ulcer margins by either (1) outlining ulcer margins onto a transparency placed directly over the ulcer (transparency method)[1-5] or (2) making a slide of the ulcer and then outlining ulcer margins from the projected slide image photographic method).[5-9] Tracings generated by either method can be quantified to yield WSA by using planimetry[l,6,7,10,11] or computer digitization.[5,8,9,11] Reliability has been reported for WSA measurements from tracings derived using the photographic method[9] and the transparency method,[11,12] but comparative reliability of measurements obtained with photographic and transparency methods has not been assessed.

Relative merits and deficiencies of photographic and transparency methods of obtaining tracings have been discussed.[1,5,6,9,10] Delineation of wound margins may be enhanced by color photography, whereas wound margins can be obscured by placement of a transparency over the wound, particularly if area lighting is poor.[9,10] Although conceptually a more "accurate" representation of wound borders might be obtained using the photographic method, scaling errors may be introduced during the process of obtaining a tracing from a slide.[6] Both photographic and transparency measurements may be influenced by the distortion involved in converting a three-dimensional multiplanar lesion located on the human body to a two-dimensional uniplanar tracing on a sheet of paper.[5,6] Comparison of photographic and transparency methods has rarely been reported. Thomas and Wysocki[5] reported that although a high correlation (r=.96) between transparency-and photographic-based measurements of the same pressure ulcers was demonstrated, WSA measurements derived from the photographic method were significantly smaller than measurements derived from the transparency method. Reasons for differences in WSA magnitude between photographic- and transparency-based methods were not discussed.

In a recent report concerning efficacy of a specific intervention for healing of pressure ulcers, Griffin et al[9] used measurements obtained by the photographic method. As an ancillary part of that study protocol, measurements were also collected using the transparency method. The authors' intent was to compare measurements obtained using the two methods, hypothesizing that if one method were "superior" to the other, differences between methods would be evident in reliability or size of WSA measurements. First, reliability of measurements obtained with each method can be estimated by the intraclass correlation. If reliabilities differ, the estimates of intraclass correlation win differ, with the less reliable method having a smaller intraclass correlation. Second, the mean size of ulcers estimated by the two methods may be compared. This test will indicate whether both methods yield measurements of WSA that (1) do not differ statistically and (2) are close to the same value (ie, are precise). Precision is defined as closeness of repeated measurements to the same value. Unless there is an unidentified source of bias, precision should be related to accuracy of WSA and therefore to validity.[13] A final goal was to determine whether precision of measurement was improved for either method by using the average of three tracings versus single tracings. Generating and analyzing multiple tracings can be time-consuming, particularly for the photographic method. Griffin and colleagues believed a study comparing photographic and transparency methods would yield information helpful to practitioners weighing the advantages of these two measurement techniques.

Specific purposes of this study were (1) to determine whether test-retest reliability was similar for WSA measurements obtained by photographic and transparency methods, (2) to compare the reliability and precision of measurements obtained using the average of three tracings versus single tracings, (3) to determine the correlation of WSA measurements obtained by photographic and transparency methods, (4) to ascertain whether mean WSA measurements obtained with photographic and transparency methods were significantly different, and (5) to analyze the relationship between WSA measurements obtained with the two methods over time. Our expectations were that test-retest reliability of measurements obtained with the two methods would be comparable and that measurements derived from photographic and transparency methods would be highly correlated, but that the magnitude of measurements yielded by the two methods might be significantly different.

Method

Subjects

Twenty inpatients (18 male, 2 female), aged 31[+ or -]16 years (X[+ or -]SD), at the Baptist Memorial Hospital Regional Rehabilitation Center (Memphis, Tenn) participated in the study. A total of 22 ulcers were studied (2 patients had 2 ulcers). All ulcers were located in the pelvic region (8 gluteal/ischial, 12 sacral/coccygeal, 2 trochanteric). Ulcers varied in size (688[+ or -]228 [mm.sup.2]) (X[+ or -]SE) and duration (13[+ or -]26 weeks) (X[+ or -]SD). Severity of ulcer grades[9] also varied, as follows: grade II=4 ulcers, grade III=12 ulcers, grade IV=4 ulcers, and grade V=2 ulcers. All participating patients or their legal guardian signed an informed consent form.

Photographic Method

At each ulcer measurement session, three photographs were taken using 35-mm color slide film(*) and an Olympus OM-2s camera.(dagger) A metric ruler was taped adjacent to and in the plane of the ulcer prior to photography. The distance between ulcer and camera was 27.9 to 30.5 cm (11-12 in). To obtain a tracing from each slide, the slide was projected onto paper, and the projector-to-paper distance and focus were adjusted until the image of the metric ruler in the slide exactly matched the ruler used in the original photograph, as described by Bulstrode et al[6] and Myers and Cherry.[7] The ulcer margins were then traced with a pen. One tracing was generated from each of the three slides for each measurement session of each individual ulcer.

Transparency Method

A transparency(double dagger) was placed directly over the ulcer, and the wound margins were traced with an indelible pen. Three tracings of the ulcer were made at each measurement session. Each transparency was washed with antiseptic soap, rinsed, and dried immediately prior to being used for ulcer tracing, as described by Thomas and Wysocki.[5] Examples of tracings obtained by photographic and transparency methods for one ulcer are shown in the Figure.

Measurement of Tracings

Tracings generated by the photographic and transparency methods were digitized as previously described[9] to obtain WSA measurements (in square millimeters). Each tracing was outlined using the stylus of a tablet digitizer(section) interfaced with an IBM-PC/XT,(perpendicular) and the area was calculated using a software program.(#) Tracings were digitized in random order.

Procedure

To assess test-retest reliability of WSA measurements, 5 ulcers were assessed using both photographic and transparency methods, and assessments were repeated after 1 hour. To compare the WSA measurements obtained by the photographic and transparency methods, all 22 ulcers were measured on a single occasion using each method. To compare photographic and transparency WSA measurements over time, the 16 ulcers that were available for longitudinal study were measured at 5-day intervals for 20 days, using each method.

Photographic and transparency methods were conducted in random order. An experienced physical therapist (JWG) performed photographic and transparency assessments and conducted digitizer measurements of tracings.

Data Analysis

Means and standard deviations were calculated using standard statistical methods. To determine intratester reliability of WSA test-retest measurements for each method, intraclass correlation coefficients (ICC[1,1])[14] were determined. A one-way random-effects analysis of variance (ANOVA) was used to obtain variance components for calculating the ICC, which was defined as the ratio of variance among patients to total variance.[14] Standard deviations for the ICCs were also calculated.[15,16] To examine further the test-retest consistency of measurements for each method, the standard error of measurement (SEM) was calculated. To compare reliability for a single measurement versus the average of three measurements, the ICC and the SEM were calculated for each method using a single measurement (randomly selected from the three measurements) and using the mean of three measurements.

To assess relationships between measurements obtained with the photographic and transparency methods, Pearson Product-Moment Correlation Coefficients (r) were computed. The null hypothesis was that no correlation existed between methods. Coefficients of determination ([r.sup.2]) were calculated to describe the percentage of variance shared by the methods. To determine the significance of the difference between photographic and transparency measurements for the 22 ulcers, a two-way mixed-model ANOVA was used, with patients and methods as the main effects. The null hypothesis was that the WSA means of the two methods were equal. To determine whether WSA measurements of the two methods differed significantly over time, a three-way mixed-model ANOVA was used, with patients, methods, and days as main effects. Correlational and ANOVA procedures were conducted, using the average of three digitized measurements for each method.

All statistical analyses were conducted using the Statistical Analysis System (SAS)(**) available through The University of Tennessee VAX computer system. The level of significance used for statistical tests was .05.

Results

The results of the test-retest reliability analyses for each method are presented in Table 1. The ICC for both slide and transparency methods was .99 using the average of three measurements. When single measurements were used, ICC values were slightly lower but not statistically different. The SEMs were larger for both methods when single measurements were used (Tab. 1).

[TABULAR DATA OMITTED]

Photographic and transparency method WSA measurements for the 22 ulcers are presented in Table 2. The Pearson correlation coefficient of .99 between WSA measurements of the two methods was statistically significant (P<.001), and the null hypothesis was rejected. The two-way ANOVA (Tab. 3) revealed that statistically significant differences existed among patients (as expected) but not between methods. Thus, the null hypothesis of no difference between means of the two methods was not rejected.

[TABULAR DATA OMITTED]

Table 3. Analysis of Variance Results for Mean Wound Surface area pf 22 Ulcers
Comparing Photographic and Transparency Methods
Source            df   SS            F        P
Ulcers            21   47878227.87   555.35   .0001
Methods            1        103.24     0.03   .8755
Error             21      86212.02
Corrected total   43   47964543.14

Table 4 summarizes the photographic and transparency method WSA measurements obtained at 5-day intervals for the 16 ulcers followed over time. Four ulcers healed during the 20-day follow-up period. Measurements from the photographic and transparency methods were significantly correlated at each time period (r=.996-.999, P<.001). The three-way ANOVA (Tab. 5) indicated that significant differences were not present for method over time, and the null hypothesis of no difference between means of measurements obtained with the photographic and transparency methods over time was not rejected.

[TABULAR DATA OMITTED]

Table 5. Analysis of Variance Results for Mean Wound Surface Area of Ulcers in
16 Patients at 5, 10, 15, and 20 Days, Comparing Photographic and Transparency
Methods.
Source             df   SS            F      P
Patients           15   15900410.69   7.09   .0001
Methods             1         19.21   0.00   .9910
Days                3    2489599.38   5.52   .0016
Methodsxdays        3       3700.29   0.01   .9990
Error              91   13677189.36
Corrected total   113   30980080.95

Discussion

A high degree of intrarater reliability was demonstrated for repeated measurements obtained with both photographic and transparency methods. The ICC value of .99 indicated that nearly all of the variance in WSA was among patients, with little variance between test and retest. The ICC calculated by using the mean of three measurements was not significantly higher than that calculated from a single measurement, for either method. Values for the SEM indicated error associated with single measurements was slightly higher than that for average measurements. The lowest ICC (ICC=.995) was found for using single measurements with the photographic method. We concluded that the photographic and transparency methods, as described in this study, provided equivalently reliable WSA measurements, whether the mean of three measurements or single measurements were used.

In our study, we determined which method yielded measurements that were closest by examining the pooled within-method standard deviations (Tab. 1). Precision was slightly improved by using the average of three measurements, particularly for the photographic method. The additional time and expense involved in analyzing three tmcings versus a single tracing, however, might not be justified for such a small increase in precision.

Photographic and transparency method WSA measurements were significantly correlated in the 22 ulcers, with the coefficient of determination indicating that the two methods had a common variance of 99% and that only 1% of the measurement variability was not shared by the two methods (Tab. 2). This finding was consistent with our expectations and agrees with the report of Thomas and Wysocki,[5] who reported a correlation of .96 between photographic and transparency method measurements of pressure ulcers located below the umbilicus. We also found that the significant correlation between methods was consistent, as healing occurred in the 16 ulcers followed over time.

Contrary to our original hypothesis, the results indicated that the WSA measurements obtained from photographic and transparency tracings did not differ significantly. Furthermore, as photographic and transparency method measurements did not differ at the 5-, 10-, 15-, and 20-day follow-up periods, the two methods continued to yield equivalent measurements as ulcers became smaller. The finding that WSA measurements did not differ significantly between the two methods disagrees with the report of Thomas and Wysocki,[5] who found that the photographic method yielded significantly smafler WSA measurements than did the transparency method. Those investigators used a different system for generating tracings from slides, using a camera system containing a reproduction ratio imptinting feature. We followed the method described by Bulstrode et al[6] and Myers and Cherry[7] in which a ruler in the photograph ensured that the projected slide image corresponded 1:1 to the original ulcer. We conclude that this protocol for generating a tracing of ulcer margins from a slide yielded an image equivalent in size to that obtained by tracing the ulcer directly onto a transparency.

The relative accuracy of the transparency and photographic methods could not be assessed because measurement of the "absolute" size of human skin ulcers in vivo was not feasible. Neither method provides a measurement of ulcer depth. A truly accurate ulcer measurement would entail quantification of skin defects in three dimensions.[6] Bulstrode et al[6] described a stereo camera/computer system for obtaining such three-dimensional wound measurements and compared these stereophotogrammetric measurements with measurements derived by transparency and photographic methods. When measuring artificial ulcers of known area, accuracy and precision were significantly better with stereophotogrammetry than with either the transparency or the photographic method. In a study of chronic in vivo leg ulcers, accuracy and precision of transparency and photographic methods were similar, although both were inferior to stereophotogrammetry.[6]

The findings of Bulstrode et al[6] indicate that photographic and transparency methods may lack accuracy. We believe the results of our study, however, indicate that both photographic and transparency methods can provide reliable measurements that can be used as indexes of healing rate for ulcers. Further research concerning such sophisticated measurement techniques as stereophotogrammetry is needed to assess the practicality and value of these measurements in patient care and clinical research environments.

We did not find evidence supporting the superiority of one method over the other for measurement of wound size, because the methods appeared to provide equivalent information. A decision between photographic and transparency methods for a particular clinical situation might therefore be based on economic aspects or patient safety concerns. In terms of time requirements and cost of material, the transparency method was more economical than the photographic method. Significantly more time was required in the photographic method, because slides had to be projected at the correct magnification and then traced prior to digitizing and because film had to be developed before a tracing could be generated. The cost for photographic method equipment was greater, because the camera (approximately $850), film, and film-processing expenses exceeded the cost for transparencies (approximately $36 for a box of 25 transparencies), antiseptic soap, and indelible pen. Either method would require a computer/digitizer system or planimeter to obtain WSA (in square millimeters) from tracings.

An advantage of the photographic method was that measurement devices did not come in direct contact with the ulcer, a desirable feature when potential contamination and tissue damage are of concern.[6] In our study, no complication was encountered by using the transparency method, involving direct contact between transparency and wound. Only six ulcers in our study, however, were grade IV or higher, and none were grossly infected. In cases of deep infected wounds, the photographic method might be the method of choice.

Interrater reliability was not investigated in this study. Because the wounds in this study were limited to pressure ulcers in the pelvic region of patients with spinal cord injury, findings might not be representative of wounds having different etiologies or different body locations. Further clinical investigation is advocated to clarify the best clinical measurement techniques for determining size changes of ulcers of different etiologies in response to treatment.

Conclusions

Photographic and transparency methods, as described in this study, provided measurements of WSA in pressure ulcers that demonstrated essentially equivalent reliability. Precision of measurement was slightly improved by using the mean of three measurements as opposed to single measurements. A significant correlation existed between WSA measurements obtained with the photographic and transparency methods. Magnitude of measurements did not significantly differ between methods, either in the 22 ulcers measured on one occasion or in the 16 ulcers measured over a 20-day period. We concluded either method can provide reliable measurements of WSA in pressure ulcers and that the two methods yielded equivalent WSA values. The transparency method was more economical than the photographic method in terms of time and equipment requirements.

(*) Kodachrome ASA-25, Eastman Kodak Co, 343 State St, Rochester, NY 14650.

(dagger) Olympus Optical Co, Tokyo, Japan.

(double dagger) 3M PP2500 transparency film, 3M Visual Systems Div, Austin, TX 98726-4599

(section) Model HDG 1111C, Hitachi Seiko Ltd, Tokyo 101, Japan.

(perpendicular) International Business Machines Corp, Old Orchard Rd, Armonk, NY 10504.

(#) CADD 3.0, Generic Software Inc, Redmond, WA 98052.

(**) SAS Institute Inc, PO Box 8000, Cary, NC 27511.

References

[1] Bohannon RW, Pfaller BA. Documentation of wound surface area from tracings of wound perimeters. Phys Ther. 1983;63:1622-1624. [2] Fergusson AG, Logan JC. Leg ulcers: assessment of response to certain topical medicaments. BMJ, 1961;1:871-874. [3] Kloth LC, Feedar A. Acceleration of wound healing with high voltage, monophasic, pulsed current. Phys Ther. 1988;68:503-508. [4] McCulloch JM, Kloth LC. Evaluation of patients with open wounds. In: Kloth LC, McCulloch JM, Feedar JA, eds. Wound Healing. Alternatives in Management. Philadelphia, Pa: FA Davis Co; 1990:97-118. [5] Thomas AC, Wysocki AB. The healing wound: a comparison of three clinically useful methods of measurement. Decubitus. 1990;3: 18-25. [6] Bulstrode CJK, Goode AW, Scott PJ. Stereophotogrammetry for measuring rates of cutaneous healing: a comparison with conventional techniques. Clin Sci. 1986;71:437-443. [7] Myers MB, Cherry G. Zinc and the healing of chronic leg ulcers, Am J Surg. 1970; 120: 77-81. [8] Akers TK, Gabrielson AL. The effect of high voltage galvanic stimulation on the rate of healing of decubitus ulcers. Biomed Sci Instrum. 1984;20:99-100. [9] Griffin JW, Tooms RE, Mendius RA, et al. Efficacy of high voltage pulsed current for healing of pressure ulcers in patients with spinal cord injury. Phys Ther. 1991;71:433-444. [10] Ramirez AT, Soroff HS, Schwartz MS, et al. Experimental wound healing in man. Surg Gynecol Obstet. 1969;128:283-293. [11] Majeske C. Reliability of wound surface area measurements. Phys Ther. 1992;72: 138-141. [12] Diamond JE, Mueller MJ, Delitto A, Sinacore DR. Reliability of a diabetic foot evaluation. Phys Ther. 1989;69:797-802. [13] Sokal RR, Rohlf FJ. Biometry. New York, NY: WH Freeman & Co Publishers; 1981:13. [14] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428. [15] Graybill FA, Martin F, Godfrey G. Confidence intervals for variance ratios specifying genetic variability. Biometrics. 1956;12:99. [16] Swiger LA, Harven WH, Everson DO, Gregory KE. The variances of intraclass correlation involving groups with only one observation. Biometrics. 1964;20:818.