Table 1.

Overview of two approaches to develop and use patient profiles

	Quantitative approach (PROFILe project)	Mixed-method approach (Tailored Healthcare project)
Objective	To develop, validate and test patient profiles as an instrument to support more tailored type 2 diabetes management in primary care	To define and validate patient profiles and to test the effect of integrating profiles in healthcare services, materials and systems on total joint replacement patients’ satisfaction with care provision
Patient profile development
Target population	Adult patients with type 2 diabetes treated in primary care	Older adults undergoing lower limb joint replacement surgery
Identification of subgroups	Growth mixture modelling	K-means clustering
Population size	~ 10,000 (development cohort) ~ 3000 (validation cohort	~ 200 (retrospective cohort) ~ 30 (qualitative interviews)
Prediction of subgroups	Machine learning	Recursive partitioning
Patient profile use in practice
Assessment: which patient characteristics are assessed?	Body mass index	Coping style
	Glycated haemoglobin	Anxiety
	Triglycerides	Communication preferences
Stratification: how are patients stratified into subgroups?	Healthcare provider enters patients BMI, HbA1c and triglycerides levels into a tool, which enables him/her to view the related subgroup with a similar glycaemic control trajectory	Healthcare provider enters the patient’s scores as determined during the consultation in a decision tree. Alternatively, patients fill out a self-reported questionnaire which is scored according to the decision tree decision rules. A suggestion for the patient’s subgroup is provided along with the level of certainty
Tailoring: how is care tailored?	Daily diabetes care planning, lifestyle information, help taking medication, frequency of consultations and emotional support are tailored according to the preferences per subgroup	Preoperative education materials and supportive systems for postoperative (tele)rehabilitation are tailored to the preferences per subgroup