Last updated: July 2001
Keywords: fungating wounds; dressings evaluation; TELER methodology; palliative care.
Due to the uncontrolled nature of fungating malignant wounds, the experiences of individual patients are unique. Experiences therefore at an individual and group level are best studied using a multiple case study approach.
Objective measurement of fungating wounds that are complex and extensive is not possible and fails to provide an adequate evaluation of palliative wound care.
A system of clinical notetaking was identified. This allows the clinician to define treatment goals and record outcomes at the individual and group level.
Indicators can be developed to measure specific parameters of dressing performance and optimise palliative wound care
This article describes the development of a methodology for researching fungating wounds. Measures were identified to optimise practice in palliative wound care. For example:
Exudate leakage, including peri-wound skin condition, dressing fit, frequency of dressing interventions and visual evidence of soiling
Dressing adherence including pain, trauma and bleeding on removal
The impact of a dressing change from the patient's perspective.
These measures were included in a system of clinical notetaking, which also recorded the key aspects of total patient care. The system adopted was based on the TELER® methodology for treatment evaluation. This is a patient-centered method of data collection and uses clinical indicators to measure outcomes.
The incidence of fungating malignant wounds is not recorded in population-based cancer registries and is therefore unknown.They are caused by infiltration of the skin by a local tumour and recurrent or metastatic spread from a primary tumour. Unless the tumour is controlled by single or combination anti-cancer treatments, the fungation extends and can cause massive destruction to the skin and adjacent structures. The nature of the damage includes proliferative growth, loss of vascularity and necrotic ulceration and can extend to sinus and fistula formation in organs such as the bladder, bowel and vagina. The significance of such wounds may be appropriately judged in terms of the impact of an uncontrolled fungating tumour on the individual, the family and health service resources over a potentially lengthy period.
The mainstays of the palliative management of uncontrolled fungating tumours include symptom control measures, both systemic and topical, and wound dressings. Palliative adjuvant therapies are also necessary for problems such as lymphoedema, which may require surgery and radiotherapy to lymph nodes ,,.
In view of a lack of research literature on the management of fungating wounds, a study was undertaken  with three significant aims:
To explore individual experiences of fungating malignant wounds.
To establish how selected dressing materials control symptoms, such as wound exudate, and reduce the impact of fungating wounds on daily life.
To introduce novel dressing systems generated from individual needs and to study the effects of symptom control and impact of the wound on daily life.
The sample comprised 45 individuals with fungating malignant wounds; the location, size, stage of development and impact of the wounds varied. The study design was longitudinal. Data sampling included participant observation of individual experiences, particularly dressing changes, and was adapted to individual circumstances and needs.
A case study approach provided an insight into the highly personal and unique experiences of individuals living with a fungating wound ,. Since common symptoms and problems have been identified by published case reports ,, it was anticipated that the systematic study of individual cases could lead to a general understanding of the problems and the development of solutions. It was assumed from the outset of the study that across a sample of cases there would be problems of a general nature concerning recurring symptoms and dressing performance. The intention was to identify patterns from the multiple case data, without losing sight of the individual nature of the experiences studied.
The choice of a case study design raised a key question concerning the nature of knowledge generated from individual case studies and whether it is legitimate to make general statements from this information ,. Support was found for the use of a case study approach to develop knowledge at the individual and group level, with the proviso that it is not always appropriate to generalise from or to every individual experience, and that generalisations 'out of context' are not legitimate ,,.
The initial methodology adopted reflected the predominance of the experimental design in wound care research in particular and clinical evaluation studies in general . This included a structured case study, modelled on the longitudinal interrupted quasi-experimental design described by Cook and Campbell . However, this proved unworkable when applied to the study of patients with fungating disease. The methodology that emerged was a form of theory-driven evaluation, prompted and steered by the methods of data collection which were revised when a conventional approach to measurement and assessment failed to capture the experiences and concerns of the study participants.
The individual, personal nature of the first study aim invited qualitative methods using participant observations and interviews. The external and apparently measurable nature of the second and third study aims suggested robust, quantitative methods of measurement. The first version of the research protocol, therefore, included an attempt to classify fungating wounds according to conventional approaches using depth of tissue loss and size together with nominal data to record symptoms and problems . The following methods were first proposed:
A wound assessment tool (the collection of nominal data, for example, odour Yes/No)
Ultrasound imaging (to capture wound depth)
Photographs (to capture wound surface measurements)
Weight of dressings (to determine exudate volume)
However, early piloting of the wound assessment tool and the wound measurement methods identified the limitations of what could be measured objectively. It had been anticipated that ultrasound imaging would determine wound volume and volume of necrotic tissue, but the imaging systems were unable to penetrate the depth of damaged tissue to produce an accurate image. Ultrasound scanning, as a method of assessing fungating wounds, was therefore rejected on the grounds that the intrusion to the patient was not justified relative to the quality of information obtained . This forced a fundamental review of the methodology and bought into question the purpose of measurement.
Without ultrasound images it was difficult to measure wound volume accurately. Image analysis of the photographs offered, at best, an inaccurate index of wound deterioration. The complexity of the wounds and the individual circumstances meant that pre-determined, formalised methods of data collection were unable to reflect the problems described by the study participants and observed by the researcher. This was highlighted by an overwhelming problem with the fit and fixation of dressings. Dressing fit involved multiple-product use in an attempt to fill and cover the wounds. This contributed to a number of confounding variables and minimal control over the interventions. Since each dressing application was another experiment on the patient, the quasi-experimental design initially proposed was not achievable, even on an n=1 basis. The patients included in the study were, in the main, treatment failures and were therefore 'outliers' to the distribution of a given sample of cancer patients . This called into question the use of methods which aim to reveal trends, central tendencies and stable statistical variances.
The key conclusions from the early pilot of the methods of data collection were that objective assessment was not achievable and that individual concerns and experiences were not represented by the approach adopted. Since the aim of the study was to reveal knowledge of fungating wound management at the individual and group level, a systematic method of collecting data for assessing dressing performance and symptom control was required.
The literature on outcome measures provided a solution to the measurement problem in the form of the TELER® (Treatment Evaluation by A Le Roux's Method) system of clinical notetaking ,,. This system facilitates the definition of treatment goals and records outcomes at the individual and group level. Following the adoption of the TELER system the methods of data collection comprised:
TELER system of notetaking including TELER indicators (to record dressing performance and other measures of optimal palliative wound care)
Medical and nursing records (to record concurrent treatments and progress)
Participant observations and discussions (to observe and explore patient experiences)
Photographs (for illustrative purposes).
Through this approach, data can be collected on the multiple variables that may impact on the outcomes of dressing performance, which may then be explained.
The TELER system provides two types of clinical information:
TELER evaluation shows whether individual patient problems are changing while undergoing an intervention.
TELER attribution is a procedure for determining whether an observed pattern of change, or lack of change, is attributable to an intervention.
The TELER indicator is an ordinal scale for measuring outcomes of patient care. The codes of the indicator mark progress or lack of progress for specific treatment or care objectives. Code 0 indicates a deficit to be overcome, avoided or delayed; codes 4 and 5 are achievable, desired goals. The reference points of the indicator trace change over a continuum between a patient deficit and no deficit by demonstrating a series of changes. The indicators have six reference points (0-5) and measure five improvements or deteriorations .
The TELER indicators focus on the control of symptoms and dressing performance  and are designed to capture individual experiences and to measure observable features of the wounds such as necrotic tissue, bleeding and peri-wound skin condition. (Examples of the indicators are given in Appendix A.) The definitions of the codes were constructed from patient and nurse reports of wound-related problems, together with participant observations ,. The indicators incorporate both theoretical knowledge of wound management, supported by moist wound healing theory, and experiential knowledge at the clinical and personal level.
The indicators were validated over a three-year period using a consensus process. Modifications were made to meet all three criteria of validity (content, concurrent and construct).
Content validity is the extent to which the indicator codes state the clinical problem and trace changes in the presentation of the problem.
The key issues for content validity were whether the definitions of the codes were accurate and able to trace the clinical changes observed. The codes of the indicators were modified when the definitions of the codes did not reflect the problems experienced. For example Code 5 for discomfort from dressings (see Indicator 1, Appendix A) was initially defined as:
5 = Able to ignore dressings
Participants, however, indicated that they were unable to ignore the dressings completely and concluded that, 'at best', they were always aware of them. The code was therefore redefined as follows:
5 = Aware of dressings but no discomfort
Further indicators for 'dressing performance', for example dressing fit, were developed when a single definition of the problem did not encompass the various ways in which the problem presented. Dressing fit is therefore measured by three indicators:
A requirement to use multiple layers of bulky dressings at 0, while a low profile dressing is preferred at 5.
Problems of fitting a dressing, sealed on all four sides, to a site such as the axilla or groin.
Problems of fitting a dressing to a site such as the head and neck.
The dressing fit issue for all three indicators reflects the amount of work involved on the part of the patient, nurse or carer, in maintaining a well-fitting dressing.
Further changes were made when indicators measuring more than one dimension of a problem were split because the assumptions underlying the physical presentation of the problem were not valid. For example, it was assumed that erythema is the first sign of skin damage. This assumption reflects the categorisation of non-blanching erythema as the accepted first sign of pressure damage. The skin is not visibly broken at this stage but reacts to injury by becoming inflamed . The multiple problems observed with the peri-wound skin on the fungating wounds studied did not however follow a hierarchy that could be described in a single format - that of 'skin damage'. Thus, the following problems with peri-wound skin needed to be differentiated: erythema due to maceration; skin breaks due to stripping by tapes and films; and irritant dermatitis. These problems also had to be distinguished from tumour infiltration, radiation damage and cellulitis. These conditions were monitored in a table entitled Skin condition (see Appendix A). Specific problems relating to skin damage from dressings were followed using the appropriate indicator.
The indicator for bleeding was also modified to distinguish between spontaneous bleeding (tumour activity), bleeding on removal (adherence and mechanical friction from dressing materials and technique), and bleeding with cleaning (friction from materials and techniques).
Finally, indicators for soreness and irritation were split to differentiate between intensity and duration of symptoms. A fundamental criterion of the measurement tool design is that only one dimension should be measured at a time . For example, if two dimensions are included there is a measurement problem when one dimension changes but the second does not change commensurately. The problem of combining two dimensions was clearly revealed when interventions to reduce soreness and irritation, for example topical diamorphine and TENS (transcutaneous electrical nerve stimulation), reduced the intensity of the symptoms but did not alter the duration commensurately.
Concurrent validity is the extent to which the indicator codes define the clinical problem in terms that are acceptable and useful to the patients and clinicians.
Concurrent validity of the indicators was addressed by modifying descriptions of the clinical problems considered unacceptable by the participants and nurses. For example, code 0 for odour was specified as:
0 = Overwhelming odour in the house/clinic/ward
Participant 024 and his wife took exception to the word 'overwhelming' and this was withdrawn on the grounds that the word was insensitive and did not add to the specification of the problem (see Indicator 8, Appendix A).
Concurrent validity, in terms of widespread applicability of the indicators by other clinicians, needs to be established beyond this study. The TELER system and the indicators may only be used under a TELER licence, which protects copyright and also gives the user access to support and advice in the use of the system and the methodology (see contact address).
Construct validity is the extent to which the differences between the definitions of the codes for the clinical problem are defined in terms that are clinically significant.
The key issue for construct validity was whether the indicators were sensitive to clinically significant change. Early versions attempted to measure intrusion of dressing changes and the ability to achieve personal goals. These were abandoned on the grounds that the participants developed significant coping strategies to minimise intrusion and achieve personal goals, or the disease itself was the limiting factor. Participant 015 for example got up at 7.00am to meet her own dressing and hygiene needs, to join her husband for breakfast at 9.00am. She did not therefore record an inability to meet needs or goals. Participant 012 stated that she was limited primarily by a mobility problem caused by tumour infiltration of the muscles of her thigh, not by the poor performance of the dressings in managing wound exudate.
As new problems emerged additional indicators were added. For example in the final stages of data collection, participant 062 developed irritant dermatitis after 17 months of repeated use of polyurethane film. The film had left residues on the skin and set up an irritant reaction. Cleansing with a special soap, provided by the manufacturer of the film (Advanced Medical Solutions - Cheshire UK), relieved the problem and raised awareness of the need to include a specific cleansing routine with extended use of this type of product. An indicator was developed to define irritant dermatitis and distinguish it from maceration or skin stripping.
In addition, a more precise definition of an existing problem was required when dressing fit did not take account of the fit of the primary dressing. The introduction of a flat sheet hydrogel and alginate foam sheet demonstrated that conformability of these primary dressings to the wound was critical to the control of exudate. Without precise fit the exudate leaked and caused maceration under the fixation film. An indicator to evaluate the conformability of the dressing to the wound was therefore added (see Indicator 6, Appendix A).
Key issues for the reliability of the TELER indicators included accurate document layout to reduce errors and assessor training . It was noted, for instance, that codes could be incorrectly entered, particularly codes 0 and 5 due to the fact that the numbering system on the descriptions of the indicators ran 0 to 5, but 5 to O on the data sheet. Although an unexpected code 0 or 5 was visible and could be challenged, the nurse who regularly attended participant 042 asked for the format to be changed to make the numbering systems consistent: 0 to 5.
Training in assessment and data sampling was necessary to avoid coding a changing outcome for problems, such as the presence of necrotic tissue and odour, when there was no evidence that a clinically significant change had occurred. One nurse described this as 'wishful thinking' and bias towards a novel product.
When the participants or relatives undertook the data collection there were also problems with reliability. Participant 024, for example, tended to underscore the exudate leakage problem (see Indicator 7, Appendix A) and acknowledged this when the researcher suggested he was being somewhat generous. With participants 047 and 087 close relatives collected the data. As the condition of each of the participants deteriorated the coding became mechanistic and unreliable as the level of direct care, and distress, increased.
Problems also arose when a real clinical gain was perceived, but the deficit in the dressing intervention was not resolved in the terms stated by the indicator. Continued reference to the deficit appeared to undermine the situation for the participant and the nurses. For example, the codes for leakage for participant 070 changed from 0 to 3 following the introduction of novel alginate and film dressings, reflecting slight soiling from the wound at the axilla. She felt she had gained significantly and told the researcher very pointedly that the slight soiling was 'nothing'. It was agreed that the problem of fitting a sealed dressing to the axilla was one for the researcher to take up with the manufacturers, but not one to keep enquiring about. The indicator code at 3 was reliable and valid but did not capture the clinical gain to this participant. An alternative approach, not attempted in this study, would be to develop a set of 'quiz-type' indicators that measure personal gains alongside dressing performance indicators .
For the purposes of data verification and analysis a qualitative approach  was adopted. Evidence was generated in the form of explanations of the patterns of dressing performance and symptom control over time. A system of reasoning was used to ensure data analysis was rigorous and followed an established procedure. The strategy involved a five stage process, moving from explanations of the individual cases to theoretical explanations of the participants' experiences in general, and to the performance of wound dressings in particular ,,,. At the end of the five stages of analysis the following qualification to the theory of moist wound healing was proposed, to accommodate exuding fungating lesions :
"Exuding fungating wounds are managed optimally through the maintenance of humidity, and therefore moisture, at the wound and dressing interface together with absorptive capacity and controlled venting to remove exudate that is excess to the interface moisture requirement. . ." p 234.
The patterns of dressing performance were interpreted in the context of all the recorded treatment variables. A key issue was to determine whether, for example, a change in wound management coincided with a difference in exudate leakage and dressing changes. Additional explanations for patient outcomes, for example concurrent treatments and variations in the application of dressings by different individuals, were included to determine their impact. The techniques for analysing the TELER system data involved scrutiny of any patterns of change, together with a descriptive analysis of the dressing performance. The latter was to take into account disease progression, adjuvant treatments and individual contextual details.
This system of reasoning (a form of theory-driven evaluation) provided a strategy for strengthening the explanations derived from the consensus process. The use of reason and the process of abstracting from the participants' highly personal and individual experiences to theoretical explanations, and where possible, to established theory, also provided a legitimate method of generalising from the particular individual cases to the wider topic of fungating wound management in general ,,,.
When evaluating dressing products for chronic wound management within a clinical trial, the current convention is to apply the causal attribution of 'healing' to the performance of wound dressings. Wound assessment tools therefore focus on objective wound measurement. When this approach was used to assess the palliative management of advanced fungating malignant wounds the study failed to record any meaningful evidence.
The difficulties encountered in measuring eccentrically shaped, extensive and uncontrolled fungating lesions forced a review of what was important in terms of the impact of the wounds on the participants and the study aims. A form of theory-driven evaluation emerged alongside the adoption of a patient-centred method of data collection, the TELER system of treatment evaluation. TELER indicators were developed to measure specific parameters of dressing performance and optimal palliative wound care. Qualitative methods, essentially a consensus process, were used to validate the TELER indicators on the parameters of content, construct and concurrent validity. Using a system of reasoning to analyse the particular individual case study data, general explanations of dressing performance and symptom control were developed.
Lessons learned by the researcher from using this type of assessment are given in Appendix B and provide an insight into some of the problems encountered.
1. Dunlop R. Cancer: Palliative care. London: Springer, 1998.
2. Mortimer P. Skin problems in palliative care: medical aspects. In: Doyle D, Hanke GWC, MacDonald N, editors. Oxford Textbook of Palliative Medicine. Oxford: Oxford Medical Publications, 1997; 617-26.
3. Mortimer P, Badger C, Hall JG. Lymphoedema. In: Doyle D, Hanke GWC, MacDonald N, editors. Oxford Textbook of Palliative Medicine. Oxford: Oxford Medical Publications, 1997; 657-66.
4. Grocott P. An evaluation of the palliative management of fungating malignant wounds within a multiple case study design. Unpublished PhD thesis. King's College London: University of London. 1999.
5. Stake RE. The case-study method in social inquiry. Educational Researcher 1978; 7: 5-8.
6. Yin RK. Case Study Research: Design and methods. California: Sage Publications Inc, 1994.
7. Fairbairn K. A challenge that requires further research. Management of fungating breast lesions. Prof Nurse 1994; 9(4): 272-7.
8. Banks V, Jones V. Palliative care of a patient with terminal nasal carcinoma. J Wound Care 1993; 2(1): 14-5.
9. Cosgrove S. Faith healing. Nurs Times 1997; 93(7): 57-8, 62, 66.
10. Runyan WM. Idiographic goals and methods in the study of lives. J Personality 1983; 51(3): 413-37.
11. Guba EG, Lincoln YS. Fourth Generation Evaluation. Newbury Park: Sage Publications, 1989.
12. Lincoln YS, Guba EG. Naturalistic Inquiry. Newbury Park: Sage Publications, 1985.
13. Stake R, Trumball D. Naturalistic generalisations. Rev J Philosophy Soc Sci 1985; 7(1): 1-12.
14. Greene JC. Knowledge accumulation. Three views on the nature and role of knowledge in social sciences. In: Guba EG, editor. The Paradigm Dialog. Newbury Park: Sage Publications, 1990; 227-45.
15. Muir-Gray JAM. Evidence-based Healthcare: How to make health policy and management decisions. London: Churchill Livingstone, 1997.
16. Cook TD, Campbell DT. Quasi-experiments: interrupted time-series designs. In: Cook DT, Campbell TD, editors. Quasi-experimentation: Design and analysis issues for field settings. Chicago: Rand McNally, 1979; 207-232.
17. Reid J, Morison M. Towards a consensus: classification of pressure sores . J Wound Care 1994; 3(3): 157-60.
18. Grocott P. Evaluation of a tool used to assess the management of fungating wounds. J Wound Care 1997; 6(9): 421-4.
19. Eiser EW. On the difference between scientific and artistic approaches to qualitative research. Educational Researcher 1981; 10: 5-9.
20. Grocott P. Exudate management in fungating wounds. J Wound Care 1998; 7(9): 445-8.
21. Le Roux AA. TELER: The concept. Physiotherapy 1995; 79(11): 755-8.
22. Le Roux AA. TELER Information Pack. Sheffield: TELER, 1998.
23. Harding K, Davis MH, Dunkley P, Harden RM, et al. The Wound Programme. Dundee: Centre for Medical Education, 1992.
24. Mishel MH. Methodological studies: instrument development. In: Brink PJ, Wood MJ, editors. Advanced Design in Nursing Research. Newbury Park: California: Sage Publications, 1998; 238-84.
25. Grocott P. Palliative management of fungating malignant wounds. J Community Nursing 2000; 14(3): 31-40.
26. Mason J. Qualitative Researching. London: Sage Publications, 1996.
27. Toulmin S, Rieke R, Janik A. An Introduction to Reasoning. New York: Macmillan Publishing, 1984.
28. Bromley DB. The Case-Study Method in Psychology and Related Disciplines. Chichester: John Wiley & Sons, 1986.
29. Grocott P. The palliative management of fungating malignant wounds. J Wound Care 2000; 9(1): 4-9.
30. Grocott P, Cowley S. The palliative management of fungating malignant wounds - generalising from multiple case study data using a system of reasoning. Int J Nursing Studies 2001; In press.
Please assess the nature of any skin problem and record the appropriate letter in the code box. The skin surrounding fungating lesions may be affected by more than one of the following conditions. The purpose here is to make a differential diagnosis and record all the problems that explain the condition of the patient's skin.
|a) = intact, no redness|
|b) = infiltrated by tumour, hard, nodular|
|c) = Oedematous|
|d) = Lymphoedematous|
|e) = Radiation damaged|
|f) = Hot and red (signs of clinical infection)|
|g) = Excoriated by tapes and dressings|
|h) = Macerated by exudate/effluent|
|i) = Irritant dermatitis|
|0 = Regularly disturbed during sleeping and waking hours by discomfort from dressings|
|1 = Rarely disturbed during sleeping hours but regular discomfort from dressings during waking hours|
|2 = Periodic discomfort|
|3 = No more than two episodes of discomfort|
|4 = Occasional discomfort|
|5 = Aware of dressings but no discomfort|
|0 = Skin is a diffuse fiery red with glazed appearance|
|1 = Skin is a diffuse fiery red|
|2 = Skin has fiery red patches|
|3 = Skin has patchy reddening|
|4 = Skin has pale pink patches|
|5 = Skin appears intact|
|0 = Skin is raw|
|1 = Skin has multiple broken patches/or blistering which coalesce|
|2 = Skin has multiple isolated broken patches/or blistering|
|3 = Skin has scanty isolated broken patches|
|4 = Skin has no more than two isolated broken patches|
|5 = Skin appears intact|
|0 = Unbearable irritation|
|1 = Distressing irritation|
|2 = Unpleasant irritation|
|3 = Uncomfortable irritation|
|4 = Some irritation|
|5 = No irritation|
|0 = Black slough which is dry and leathery and covers the wound|
|1 = Brown/yellow slough in a thick layer, may include patches of black slough|
|2 = Yellow slough in a thick layer, may include thick patches of brown slough|
|3 = Yellow slough in a thin layer with isolated thick patches of brown/yellow slough|
|4 = Yellow slough in a thin layer, wound bed showing|
|5 = Clean wound bed|
|0 = Dressing(s) did not conform to site/shape of wound, one edge only attached, adjusted* between dressing change(s)|
|1 = Dressing(s) did not conform, as above, no adjustments|
|2 = Dressing(s) did not conform, two - three edges attached, adjusted between change(s)|
|3 = Dressing(s) did not conform, two - three edges attached, no adjustments|
|4 = Dressing(s) conformed, all edges attached, adjusted between change(s)|
|5 = Dressing(s) conformed with all the edges attached, no adjustments|
* adjusted/adjustment = re-taped, re-padded or re-positioned.
The descriptors indicate the degree of saturation. Sodden indicates there is absolutely no more capacity, at soaked there may be some, and at wet there is some capacity
|0 = Dressing(s) and (bed) clothes are sodden|
|1 = Dressing(s) and (bed) clothes are wet|
|2 = Dressing(s) wet and (bed) clothes are damp|
|3 = Dressing(s) wet and (bed) clothes are soiled in patches, size of a 50 p piece|
|4 = Dressing(s) only is wet|
|5 = Dressing(s) only is soiled|
|0 = Odour is obvious in the house/clinic/ward|
|1 = Odour is obvious at arms length from the patient|
|2 = Odour is obvious at less than arms length from the patient|
|3 = Odour is detected at arms length|
|4 = Odour is detected by the patient only|
|5 = No odour|
The following reflections are those of the researcher on completing the above research project into the palliative management of fungating wounds. The researcher was experienced in hospice-based palliative care and had 'hands-on' experience of managing fungating wounds. She was also practised in working with patients and families facing terminal disease and bereavement.
Flexibility of approach: The longitudinal study design meant the researcher had repeated contact with patients and families through the clinicians who had responsibility for the day-to-day care of the patients. From this, two key lessons were learned.
Approach to the research - A decision had to be made as to whether the researcher should maintain an objective distance or become a participant observer. The clinical needs indicated a problem-solving approach and a participant role. However, this meant that the initial objective approach to data collection could not be sustained. The methods therefore had to change to a qualitative design to accommodate the 'real life' setting.
Need for sensitivity and flexibility - When collecting data it was sometimes necessary to withdraw even if it meant that no data were collected.
Impact of the consensus approach: The complex and interrelated nature of problems in palliative wound care meant that single 'cause and effect' deductive analysis could not be used as part of the clinical decision-making process. The process adopted involved interpreting multiple changing signs and symptoms and reaching a consensus with the patient on the most probable cause of a number of interrelated problems. The lesson learned from this observation was that the researcher had to accept a level of uncertainty and to collect multiple sources of largely uncontrolled data to capture the problems and adopt a qualitative interpretative approach to the data analysis.
Participant observation: Participants were involved in the study for some months and in two cases for over two years. They were followed in their homes, the oncology units and hospices. Over time the participants requested that the researcher act as their advocate and accompany them to oncology outpatient visits and contribute to the consultation. On occasions, the community nurses asked the researcher to report particular problems or take a photograph or wound swab directly to the hospital with a request for urgent feedback and advice. In addition, when the researcher observed patterns of problems with individual patients, such as recurrent infection which exacerbated unpleasant symptoms of exudate and odour, she voiced these and entered into the clinical decision-making process. This aspect of participant observation led to a possible breech of the boundary between the research role and an assumption of clinical responsibilities, which should not have been part of the researcher's remit. However, the clinical needs of the participants were uppermost and there was a gradual tacit acceptance by the clinicians of this aspect of the researcher's role. Three key lessons were learned:
Patient needs take precedence over research interests. It is a matter of researching with patients and not on patients.
Uncomfortable experiences are often a part of clinical research and it is necessary to accept these feelings, which diminish with time and experience.
The research process should document all such uncomfortable experiences in the field notes or in a diary. They need to be included in a section in the research report on ethical issues arising from the study. The very act of writing and reflecting on the experiences can help to make sense of them.