Table 2.

Qualitative Data Collection Methods

Method	Sampling strategy	Number of resulting documents	Detail
Observation	Convenience Shadowed teams with multiple DM appointments in single day All relevant meetings and trainings, as allowed by clinics As possible when in clinics for meetings or interviews	126 field notes	Shadowed teams at all 11 clinics as they cared for patients with diabetes Observed relevant clinic & team meetings and trainings Informal observations & conversations throughout study
Semi- structured interviews	Purposive Sampled for high and low prescribers; MD/DO vs NP/PA; range of enthusiasm for the intervention	34 transcripts	Explored the thoughts and opinions of clinic staff as related to the implementation process and the intervention itself. Interviewed 23 PCPs (MD = 15; PA/NP = 8) and 11 RNs
Group discussions	Purposive Sampled for diversity of staff role across clinics and organizations	8 transcripts	Guided discussions that explored within-group opinions as related to the implementation process and the intervention itself. Stand-alone or dedicated time during routine staff meetings 8 separate group discussions divided by clinic role. Participation by a total of 79 staff: 27 PCPs, 16 RNs, 19 MAs, 7 TAs, 6 PCCs, 2 administrative, 2 pharmacists
Diaries by site coordinators	Not applicable	31 months of entries	Clinic-based study site coordinators (4) wrote weekly entries about the surprises, challenges, solutions, unresolved issues and day-to-day logistics of implementation based on informal observations and discussions. Monthly email exchanges between qualitative researchers and site coordinators to clarify and expand on original entries.
Document collection	Not applicable	201 documents	Relevant clinic & contextual documents (e.g., in-house newsletters, plans to implement health care reform) Communications (e.g., email strings among the study team; outreach to clinics)
Chart review	Varied by organization Org A: all patients indicated but not on an ALL medication (195 charts) Org B: purposive sample of patients indicated but not on an ALL medication from 9 providers at 5 (of 6) clinics (100 charts) Org C: List of all patients seen in past 36 months and indicated but not on in all 4 (of 4) clinics, filtered by medical record number; reviewed first 136 (136 charts)	431 unique patients	Goal: Determine why some patients considered indicated for an ALL medication (statin or ACEI/ARB) per intervention logic are not prescribed the medication. One site coordinator at each organization reviewed charts from sample of patients indicated for but not prescribed an ALL medication.

Abbreviations: PCP, Primary care provider; MD, Doctor of Medicine; RNs, Registered nurse; MAs, Medical assistant; TAs, Team assistant, PCC, Patient Care Coordinator; PA, Physician assistant; NP, Nurse practitioner