Table 1 Summary of variables

Number of visits to family physicians

Number of visits to family physicians in a 12-month period.

Measure of realized access and utilization of primary care

Discrete visits counted as ≥1 billing code, per date, family physician, and primary care location

Continuous variable

Concentration of Care Index – family physician level

\(COCI(FP)=\frac{\sum_{i=1}^k{n}_i^2-N}{N\left(N-1\right)}\)

k = total number of family physicians visited

N = total number of visits during the year

n_i = number of visits to physician i

A proxy for relational continuity of care

Different family physicians were identified using RAMQ unique physician identifier, physician speciality and type of organization to select primary care locations.

Continuous variable, ranges from 0 (all visits to different providers) to 1 (all visits to the same provider)

Concentration of Care Index – practice level

\(COCI(P)=\frac{\sum_{i=1}^k{n}_i^2-N}{N\left(N-1\right)}\)

k = total number of practices visited

N = total number of visits during the year

n_i = number of visits to practice i

A proxy for continuity of care with the practice

Different practices were identified using RAMQ unique identifier for delivery organization, unique physician identifier, physician speciality and type of organization to select primary care locations. Non-physician visits are not available in RAMQ data.

Continuous variable, ranges from 0 (all visits to different practices) to 1 (all visits to the same practice)

Pre/post attachment periods

12-month periods relative to date of attachment (T0)

Attachment as a potential enabling determinant

Categorical variable with 4 periods:

0) T0–2 years (731 to 366 days before attachment)

1) T0–1 year (365 to 1 day before attachment)

2) T0 + 1 year (Attachment date to 365 days after attachment)

3) T0 + 2 years (366 to 731 days after attachment)

Age

Age at the date of attachment, in years.

Predisposing determinant

Categorical variable: 1–5, 6–17, 18–34, 35–54, 55–69, ≥70 years old

Sex

Sex as indicated in patient information in billing data

Predisposing determinant

Dichotomous variable: male, female

Medical vulnerability

Centralized waiting lists identify a patient as medically vulnerable if they have at least one health condition among a list of 19 (e.g., diabetes, mental health problem, hypertension) or are ≥70 years old [59]. Vulnerability is assessed based on self-reported patient data upon registration, a phone evaluation by a nurse (if needed), and is verified by the family physician upon the first visit.

Need determinant

Determined using the billing codes for attaching vulnerable/non-vulnerable patients through centralized waiting lists (non-vulnerable: 19952; vulnerable: 19951 and 19,956).

Dichotomous variable: vulnerable/non-vulnerable

Charlson Comorbidity Index

Comorbidity index based on CIM-9 diagnostic codes in billing data and adjusted for age [60, 61]. Because diagnostic codes may be underreported for unattached patients facing barriers to accessing care, index was measured for each 12-month period (relative to attachment date) and the maximum value across all periods was selected for each patient.

Need determinant

Categorical variable: low (0), medium (1–3), and high (≥4) comorbidity.

Categories for this study were determined upon reviewing score distribution in the included population and bivariate analyses.

Remoteness of health region

Type of health region where patient resides. Remoteness is determined by the Ministry of Health and Social Services [62], based on population from census data, presence of large centers and influence of universities.

Enabling determinant

Categorical variable: university (urban), peripheral, intermediary or remote