Clustering of behavior related data

Dear Users,

I'm quite beginner in this field but now my research requires some methodology and I thought to create a topic, maybe somebody had the similar issue before.
I have some data regarding to health-related features, including:
- BMI (scale)
- Current diseases (categorical)
- Physical activity (scale, how long the participant does sport in a week, in hours)
- Tobacco use (scale, how often the participant smokes in a day)
- Alcohol use (scale, consumed alcoholic beverages in the past month)
- Sedentary behavior (scale, how long the participant have to sit in a day, in hours)
I know it would have been better to include variables, regarding to eating habits, which I had, but it turned out that my script failed to save that data.
In the past days I've been trying to achieve a clustering based on this data, but I'm sure that I'm missing some important steps. The aim is to use fuzzy clustering to classify the behavior in terms of risk. When I use k-means clustering "just like that", the clusters are generated based on the distances, of course, but is it possible to direct the process in order achieve the mentioned result? To define that if a specific variable is higher, then it result in higher risk for example. Or would it be a classification problem that require labelled records?
I'm using R and Python (Scikit, TensorFlow etc.) normally.
Sorry, if my question does not make sense.

Thank you very much for your help in advance!
Kind regards,