Effective pain management can significantly improve quality of life and outcomes for various types of patients (e.g. elderly, adult, young) and often requires assisted living for a significant number of people worldwide. In order to improve our understanding of patients' response to pain and needs for assisted living we need to develop adequate data processing techniques that would enable us to understand underlying interdependencies. To this purpose in this paper we develop several different algorithms that can predict the need for medically assisted living outcomes using a large database obtained as a part of the national health survey. As a part of the survey the respondents provided detailed information about general health care state, acute and chronic problems as well as personal perception of pain associated with performing two simple talks: walking on the flat surface and walking upstairs. We model the correspondent responses using multinomial random variables and propose structured deep learning models based on maximum likelihood estimation and machine learning for information fusion. For comparison purposes we also implement fully connected deep learning network and use its results as benchmark measurements. We evaluate the performance of the proposed techniques using the national survey data and split them into two parts used for training and testing. Our preliminary results indicate that the proposed models can potentially be useful in forecasting the need for medically assisted living.