Archives of Design Research

Our target subjects were new haemodialysis patients who had received haemodialysis for less than a year. Their life patterns and dietary habits should be changed a lot along with the changes from life without dialysis to life with dialysis. Therefore, they are most in need of changing their behaviour for weight management, and they will have their unique and typical behaviour that is different from other normal dieters.

We individually interviewed four participants in depth to explore and understand their unique and typical weight management behaviour. We determined that the number of four would be sufficient to extract insightful results. It is because the treatment of haemodialysis has a substantial impact on patients’ overall lifestyle, which makes their weight management patterns typical. Basically, they are provided with the guidelines about their healthcare, including diet, based on identical medical knowledge. They also have to visit the hospital regularly about three times a week for dialysis and four hours for each visit, which also greatly affects their life patterns. Based on these points, it was presumed that they have their own unique behaviour features for self-weight-management in the sense of their motivation, the root of risk perception, the difficulties, and the solutions they have found. It is clearly distinguished from those of other dieters who don’t have any serious problems. Due to the very typical nature of the patients’ weight management patterns, four interviewees (two males and two females) were deemed sufficient to represent the patients’ needs and reveal insights.

The four patients had received haemodialysis for three, five, six, and ten months, respectively. They were between their 40s and 50s, and the sex ratio was half and half (Table 1). They were recruited through a domestic web site for kidney disorder patients. In the case of P4, the mother of P4 also participated in the interview. Interviews were conducted at a place where they felt comfortable, such as their houses or cafés or the restaurant run by one of them.

Interviews were conducted to identify the hidden and underlying needs of the participants. In order to identify their deepest intrinsic needs, we used the ‘five-whys’ technique (Reid & Smyth-Renshaw, 2012). Further, a probe toolkit was used to allow them to be aware of and express their needs that are unmet, unarticulated or even unknown (Mootee, 2013). The kit was used to help invoke participants’ memories and remind them of their specific behaviours at a certain time and space. The kit consisted of two tools, addressing ‘time’ and ‘space’. The first tool was a timetable (Figure 2(a)) that reminded the participants of their personal experiences according to ‘time’. The second tool reminded them of memories related to ‘space’ based on cards with general living products and specific places in the home, such as the living room and kitchen (Figure 2(b)). Participants looked back at their usual behavioural patterns by creating a timetable and explained in detail how their house structure looked, where a scale was placed, and when they measured their weight by using space-product cards.

Figure 2 Probe Toolkit: (a) Time table (first) (b) Space-product cards (second, third)

The questions were mainly about daily life patterns and how the participants manage their weight (weight measurement and behaviour management). First, to help participants actively think about their daily life and their behaviours, probe tools were shown to them. Next, specific questions were asked to understand their underlying needs in the weight management process, how and when they measure their weight, and what they usually do to manage their behaviour.

The interview data were analysed to 1) identify the patients’ behaviours and needs concerning self-management of weight and 2) build a new behaviour process model tailored to the identified behaviour and needs of haemodialysis patients. The analysis was processed in three steps: 1) Transcription, 2) Coding, and 3) Discussion. First, the recorded interviews were transcribed into a total of approx. 9,000 Korean words. After that, the interview transcript was coded using the factors from the stages of the HAPA model as a priori coding scheme (Table 2). There were a total of eight codes, risk perception, positive outcome expectancies, intention, action plan, coping plan, action & action control, self-efficacy, and other barriers and resources.

The interview data were coded by three design researchers with more than four years of experience in user-centred design research, skilled in qualitative content analysis. The coders categorised bundles of sentences with one meaning into each code. For instance, the following sentences are one bundle of semantic units: “I check my weight once before and after a meal. I compare the weights and I see how much my weight gained for each meal”. After the data was coded individually, three coders discussed the parts that did not agree with each other to build final consensus for the final coding result and extract insights related to the health behaviour of haemodialysis patients. As a result of coding, the following numbers of semantic units composed of several sentences were collected for the total of eight codes: risk perception (17), positive outcome expectancies (1), intention (6), action plan (10), coping plan (8), action & action control (23), self-efficacy (8), and other barriers and supports (19) (table 3).

For inter-coder reliability, two Kappa coefficients between main coder and other coders were obtained. Nvivo11 was used to calculate Kappa coefficients. The average Kappa was 0.75, signifying substantial agreement (Viera & Garrett, 2005). Through the discussion among three coders, the major findings were extracted for each stage with a focus on developing a more practical and deeper understanding of the experiences of haemodialysis patients. For example, we found that the greatest motivation factor to the haemodialysis patients was risk perception rather than positive outcome expectancy. Patients' responses often refer to their own health risks that may be caused by their actions, rather than anticipating positive outcomes with their own actions. It was also found that the patients’ goals were to maintain their current status quo, whereas the goal of normal health behaviour is to make the body better, such as to lose weight or to make a beautiful body. Subsequently, a behaviour change model tailored for haemodialysis patients was constructed, and possible design implications were discussed.

It was found from the results of the coding analysis that each stage of the HAPA model (preintenders, intenders and actors) has a different function in the self-weight-management of haemodialysis patients: 1) motivation, 2) preparation and 3) action. The three functions could act as major criteria for designing an intervention in patients’ behaviour change for. Therefore, the specific features of patient behaviour which need to be considered for fulfilling each function were identified. Finally, a model for identifying the self-weight-management behaviour of haemodialysis was developed. The model consists of three main stages and the key features to be considered in each stage (Figure 3).

Figure 3 Model of the self-weight-management behavior process

In the motivation stage of the self-weight-management behaviour process, risk perception is much more effective for motivating the patient for weight management than positive outcome expectancy, since risks can lead to grave consequences. Factors that raise patients’ awareness can be used to lead patients to healthy behaviours. More detail design directions for this strategy are as follows:

ㆍ Design for users to easily recognise the danger signals of the body in daily life.

Daily body signals like a feeling of heaviness or swollen face give patients risk perception. Therefore, it would be helpful to let patients perceive their own risks with design features that would make them more sensitive to body signals in daily life. To do that, multiple body data sensing techniques could be used to detect small changes, or small changes could be visualised to make them more noticeable. One of the most basic body data sensing tools is a scale. However, the current scale only guides the weight in numbers. Patients must be aware of trivial changes in their body with a very small number of changes of less than 1 kilogram, and always have to do an additional calculation to recognise their weight changes. By simply changing the function and user interface of this scale, patients will be able to more easily recognise their body weight changes. For example, design a scale that calculates and shows the weight differences between a patient’s dry weight and his/her current weight. At the same time, it would be helpful to add visual indicators so that patients can notice their changes more easily and intuitively. The solution can be approached in a more persuasive way. In this case, the weight data also can be designed to be shown along with the patient's behaviour pattern. For example, if the weight data can be measured and shown in front of the refrigerator or in the bathroom just before patients eat or after patients go to the bathroom, the data could have a more positive impact when patients take actions or plan next actions.

ㆍ Design features that send a warning signal if the user is negligent in weight management or weight has changed excessively over a long period of time.

Although it is possible to manage weight temporarily from risk perception, risk perception and motivation would be reduced due to the repetition of behaviour patterns over a longer period of time. Therefore, it is possible to design functionalities which alert patients to be careful if they do not measure their weight for many hours or if the weight data change excessively. A mobile application could be developed to record and track patients’ body data changes and give alarms through their mobiles. There are some mobile applications for dialysis patients, but there is still lack of specific applications for self-weight-management of haemodialysis. The mobile applications also should be friendly and easy-to-use to keep them engaged.

ㆍ Design to motivate patients to manage their weight using risk experiences from other patients.

Risk can also be perceived externally through indirect experience from fellow patients or advice from physicians or the media. In particular, patients are susceptible to a great deal of influence from other patients through short conversations at a hospital, hearing other patients in conversation with medical staff, and experiences from other patients shown in the media, including websites. All interviewees mentioned their fellow patients who did not manage their weight well and stated that they received indirect risk perception from them. Therefore, it would be good if a service were designed in such a way that it could raise patients’ risk perception by providing other patients’ risk experiences. There are many ways to see and share other patients’ experience. One of the possible ways is to utilise the web community site for the dialysis patients. Although it is currently used well for sharing information and personal experience, it would be good to be utilised more actively for risk perception and behaviour change for dialysis patients.

At the preparation stage, all haemodialysis patients had a goal of limited weight gain during the inter-dialysis period or a given day. Their goal was to focus on maintaining their current weight so as not to become worse. This goal was expressed in the form of setting a weight gain threshold. The resulting actions were usually diet or exercise according to the patients’ conditions and goals. However, patients sometimes planned for unhealthy but desired behaviour, such as drinking. More detailed directions for Design Strategy 2 include:

ㆍ Design for the patients to have a good action plan for achieving their goal by reflecting their current weight status accurately.

Patients had their own ways of taking action to keep their goal of limited weight gain. They tended to take immediate actions in response to their weight status at a given time rather than establishing a long-term and continuous plan. They checked their weight by measuring weight using a scale or checking for facial swelling, hence the design examples given above. One possible way is to provide various recommended action lists based on the measured weight data, letting the user know how much they should exercise or how much food intake they can have for achieving their goal. In detail, a mobile application can be designed to continuously record patient behaviour and subsequent weight change data. In other words, when patients take a walk, the application keeps track of how much their weight is losing and can provide a personalised action guide based cumulated data.

ㆍ Design features enabling the patients to have flexible action plans considering variables such as weather and eating out.

The personal intention is important to weight management. However, the weight-management behaviour is influenced by external factors and circumstances in reality. This is because weight management is directly related to food and water intake that are easy to be influenced externally. For example, patients may drink more water when the weather is hot or eat more food when eating out with friends, which prevents them from self-regulating their desire or overcoming social pressure. Therefore, it is necessary to take into account such barriers in the design of products or services for weight-management in haemodialysis patients, with the ultimate aim of achieving the goal rather than allowing the patients to fail. The design should address recording and managing weight by including not only obvious variables, such as the amount of food and water intake, but also external factors that influence weight, such as the weather and social gatherings. A diary or a calendar can be designed for patients to record their external circumstances or their personal plans. It can help patients to remind their plans and help to plan their meals. If the diary or calendar is in web-based or mobile-based application, then it can actively interpret patients' schedules and suggests their meal plans. If patients have a plan of having dinner out, then reducing the amount of food in advance can be encouraged.