Table 3. Quality and Utilization Outcomes Associated With Reporting a Usual Cliniciana.
| Outcome | Respondents, No. (%)b | Unadjusted difference, percentage point | P value for unadjusted difference | Adjusted marginal difference, percentage point (95% CI)c | |
|---|---|---|---|---|---|
| No usual clinician (n = 1865) | Usual clinician (n = 23 625) | ||||
| Receipt of preventive care | |||||
| Blood pressure checked | 1717 (91.2) | 23 363 (98.8) | 7.6 | <.001 | 6.3 (4.6 to 7.9) |
| Cholesterol checked | 1639 (88.2) | 22 762 (96.8) | 8.5 | <.001 | 6.7 (5.4 to 8.1) |
| Diabetes screening | 556 (46.3) | 8375 (53.0) | 6.7 | .006 | 4.3 (1.2 to 7.4) |
| Colorectal cancer screening | 452 (34) | 5729 (41.2) | 7.2 | <.001 | 4.6 (1.4 to 7.9) |
| Influenza vaccine | 1130 (57.6) | 17 681 (73.4) | 15.9 | <.001 | 11.6 (9.2 to 14.0) |
| Pneumonia vaccine | 1258 (64) | 18 147 (74.7) | 10.7 | <.001 | 6.5 (3.8 to 9.2) |
| Shingles vaccine | 180 (26.2) | 2675 (36.9) | 10.8 | .02 | 6.3 (2.2 to 10.5) |
| Spending | |||||
| Outcome | No usual clinician, mean (SD), (n = 1755), $ | Usual clinician, mean (SD), (n = 21 360), $ | Difference, mean, $ | P value for difference | Mean (95% CI), $ |
| Total spending | 5066.24 (283.55) | 6592.75 (95.38) | 1526.51 | <.001 | 891.75 (15.61 to 1767.90) |
| Out-of-pocket spending | 708.96 (61.19) | 878.59 (28.56) | 169.63 | .02 | 31.87 (−158.49 to 222.23) |
Data source was Centers for Medicare & Medicaid Services, Medicare Current Beneficiary Survey, Survey File Data, 2010, 2013, 2016. Diabetes screening asked in 2010 and 2016 only; colorectal cancer screening asked in 2013 and 2016 only; shingles vaccine question asked in 2016 only.
Overall, 25 490 respondent-years were included in this analysis. All percentages are weighted.
The multivariable models for preventive care adjusted for sex, age, race and ethnicity, income, educational status, marital status, living situation, primary insurance, supplemental insurance, census area, frailty quartile, number of chronic conditions, and self-reported health and excluded 93 respondents with 1 or more missing responses. Income reported as do not know or refused to answer was treated as a separate category. The multivariable models for cost adjusted for the same covariates and excluded 2375 respondents (9%) with missing responses.