Association Between Primary Care Utilization and Amputation Risk for Patients with Peripheral Arterial Disease

Primary care plays a critical role in identifying and managing comorbidities that can impact a patient's amputation risk. The direct impact of primary care utilization on amputation risk has never been evaluated. We investigated the association between primary care exposure and primary care utilization on amputation risk among patients with peripheral arterial disease (PAD). We conducted a case-control study for patients with PAD undergoing lower extremity interventions at a tertiary health system between November 3, 2017, and November 4, 2022. Data were extracted from the electronic health record (EHR). Cases were defined as patients with PAD who underwent major lower extremity amputation, and controls were patients with PAD who underwent non-amputation lower extremity interventions (bypass, endarterectomy, or angiographic intervention). Primary care exposure was defined as EHR documentation of a primary care provider (PCP). Primary care utilization was coded as a binary variable, where patients were considered exposed if they had a PCP visit during the study period. Patients were excluded from the utilization assessment if their PCP was external to our health system to ensure accurate data capture. Multivariable logistic regression was used to control for age, sex, race, smoking status, insurance type, and median income based on zip code. Geographic variation in exposure and outcome variables was assessed. We identified 4403 patients with PAD who underwent lower extremity interventions, with 1561 (35%) undergoing amputation. A total of 4131 (94%) had a documented PCP. Compared with patients undergoing non-amputation lower extremity interventions, patients who underwent amputations were significantly younger and more likely to be male, Black, reside in a low-income zip code, have a history of diabetes, and have higher 30- and 90-day readmission rates after surgery (Table). Patients who received amputations were 52% less likely to have a documented PCP than patients with other vascular interventions (odds ratio, 0.48; P < .001). Of 917 patients with a PCP within our health system, 650 (71%) had a primary care visit during the study period and 303 (33%) underwent amputation. Among this subset, patients who underwent amputations were significantly less likely to have had at least one PCP visit than patients with other vascular interventions (odds ratio, 0.44l P < .001). Geographic variation in amputation is shown in Fig 1. Both PCP exposure and utilization were associated with lower rates of amputation in patients with PAD, even after risk-adjustment for known confounders. Further studies will be needed to establish any causal link between primary care and lower amputation risk.TableDifferences in representative covariates between patients with peripheral artery disease (PAD) who underwent major amputation vs other non-amputation lower extremity vascular proceduresAmputation (n = 1561)Non-amputation lower extremity intervention (n = 2842)P-valueAge, years65 (13)67 (11)<.001Male sex1024 (66)1752 (62).009Race<.001 White755 (48)1787 (63) Black744 (48)969 (34) Other62 (4)86 (3)Income by zip code, $61,521 (31,590)64,857 (30,595).003Surgery type<.001 Angiographic intervention0 (0)1815 (64) Endarterectomy0 (0)176 (6) Bypass0 (0)851 (30) Amputation1561 (100)0 (0)Comorbidities Diabetes744 (48)987 (35)<.001 Hemoglobin A1c, mmol/mol7497 (7.336)6848 (1.693)<.001 Hyperlipidemia270 (17)584 (21).009 Coronary artery disease168 (11)327 (12).45 Hypertension804 (52)1623 (57)<.001Smoking status<.001 Active smoker291 (19)651 (23) Former smoker680 (44)1570 (55) Never smoker538 (34)554 (19) Unknown51 (3)67 (2)Insurance type.5 Medicaid190 (12)322 (11) Medicare912 (58)1643 (58) Private459 (29)877 (31)Disabled (prior to surgery)439 (29)553 (20)<.001Primary care exposure1422 (91)2709 (95)<.001Readmissions 30-day readmissions506 (32)554 (19)<.001 90-day readmissions670 (43)896 (32)<.001Boldface P-values indicate statistically significant differences (P < .05).Data are presented as number (%) or mean (standard deviation).Age, sex, race, insurance type, smoking status, and median income were included in the multivariable models. Open table in a new tab