The SID Pain App: Design, development and testing of a pain measurement application for use in the care for adults with a severe or profound intellectual disability

Pain is a common occurrence in the life of a person with a severe or profound intellectual disability (SID). They often have co-morbid illnesses and run a higher risk of contracting infections or getting injured. Due to the severity of their disability, people with SID have trouble expressing and communicating, leading caregivers to find pain assessment one of the most challenging tasks. Pain can be missed or detected too late, which leads to a worsening of illness, hospital stays or premature death. Since self-report of pain is impossible and systematic pain observations miss subtle and idiosyncratic ways of pain expression, a system was developed to physiologically measure pain. This system consists of a smart sock measuring skin conductance, a transmitter that sends the physiological data via Bluetooth®, and the SID Pain App on a mobile device. This dissertation describes the design, development and testing of the SID Pain App, with aims to (1) evaluate several physiological pain detection methods; (2) develop a physiological pain algorithm to be programmed into a mobile application; and (3) test the SID Pain App on adults with SID and assess the user experience of the users. To evaluate several physiological pain detection methods a systematic review was conducted. From 29 reviews, describing results from 540 articles, 1,054 combinations were found of a physiological pain detection method and a target group. When vulnerability of target group and invasiveness of measurement method was graphically displayed, the trend showed that invasive methods were mostly used on the least vulnerable target groups, and vice versa. The pain detection algorithm to be placed in the SID Pain App was developed with pain data from healthy adult participants. Each participant put a hand in ice-water until pain threshold was reached. Participants self-reported when they felt no pain, discomfort, or had reached their pain threshold. This resulted in 68,900 data points of pain and no pain, which were used to train and test models with Random Forest Prediction. While the reliability of the final tree model was high, disbalance in the data caused the precision and accuracy to be around 0.50. When 50% of pain data was interpolated, which means that similar data was generated based on the test data, precision and accuracy rose to 80%. To ascertain what a pain detection application would need to be functional and useful in daily practice, the SID Pain App was designed in collaboration with caregivers and parents of adults with SID. Via several methods, wishes, needs, constraints and opinions were gathered among caregivers, which led to three distinct designs. The design that was deemed the best by caregivers was evaluated by a focus group of experts and adjusted to be placed in the SID Pain App. The usability of the SID Pain App was examined among end-users who were asked to use the app. Answers on a user experience questionnaire demonstrated that the SID Pain App had positive scores on all six aspects. On every aspect, except perspicuity, the SID Pain App scored above average compared to a benchmark of similar products. As a final step, the SID Pain App and its pain classification algorithm was tested on adults with SID during physical therapy. Participants experienced 2-4 painful moments during the session. The SID Pain app detected a painful moment 6 seconds before a pain observation was made. This study is the first study to design, develop and test a physiological pain detection method on adults with a severe or profound intellectual disability. The results are promising, because timely pain detection is very important in people who are unable to communicate about their pain.