Table 6. Findings of studies examining health systems arrangements relating to health systems delivery and governance.
| Study | Setting and Sample Size | Study Design | Findings (95% CIs Given in Brackets Where Available)ORs Are Adjusted for Confounding Unless Stated Otherwise. | Risk of Bias Assessment |
| Routine place of care for HT | ||||
| Angell et al. 2008 [56] | US. Urban population from NYCn = 1,975 | Cross-sectional | HT awareness with a routine place of care 85.1% versus 65.5% without (p<0.05). HT treatment with routine place of care 76.4% versus without 42.1% (p<0.05). OR for HT control without a routine place of care 0.21 (0.07–0.66) versus with a routine place of care | Low risk of bias. |
| He et al. 2002 [53] | US. General populationn = 4,144 | Cross-sectional | OR for control for same health facility of care 2.77 (1.88–4.09) versus lack of same facility of care | Low risk of bias. |
| Hyman and Pavlik, 2001 [58] | US. Nationally representative samplen = 10,576 | Cross-sectional | OR for lack of awareness of HT: has usual source of care: 1.12 (0.87–1.43) versus has no usual source of care. OR for acknowledged uncontrolled HT: has usual source of care: 1.07 (0.63–1.84) versus no usual source of care. | Low risk of bias. |
| Moy et al. 1995 [55] | US. Nationally representative samplen = 6,158 | Cross-sectional | OR for no HT treatment (reference 1 for physician's office) Clinic OR = 1.07 (0.90–1.28), Emergency department OR = 1.36 (0.73–2.55), No usual place of care OR = 3.94 (3.05–5.08) | High risk of non-differential misclassification. Unclear risk of differential misclassification. |
| Nguyen et al. 2011 [71] | US. Population sample from NYCn = 1,334 | Cross-sectional | HT awareness: OR = 1.0 (0.2–5.6) no usual care versus usual place of care (baseline). HT treatment OR = 0.2 (0.1–0.8) no usual care versus usual place of care (baseline). Systolic BP 16.4 mmHg higher with no usual place of care (p = 0.02). | Low risk of bias. |
| Spatz et al. 2010 [70] | US. Nationally representative samplen = 6,672 | Cross-sectional | APR for being untreated = 2.43 (1.88–2.85) for no usual source of care versus having a usual source of care. | Low risk of bias. |
| Routine physician for HT care | ||||
| Shea et al. 1992a [35] | US. Hospital-based African American and Hispanic inner-city population in NYCn = 207 | Case-control | OR for severe uncontrolled HT with no routine physician 3.5 (1.6–7.7) versus with a routine physician | Unclear risk of differential and non-differential misclassification. |
| Ahluwalia et al. 2010 [73] | US. West Virginian women in a screening initiativen = 733 | Cross-sectional | OR of having uncontrolled HT with a regular physician 0.34 (0.13–0.88) versus no regular physician | High risk of sample bias. Unclear risk of non-differential misclassification bias. |
| He et al. 2002 [53] | US. General populationn = 4,144 | Cross-sectional | OR for HT control same health provider of care 2.29 (1.74–3.02) versus lack of same provider of care | Low risk of bias. |
| Hill et al. 2002 [57] | US. Inner-city African American men presenting to the emergency departmentn = 309 | Cross-sectional | Non-significant association between regular MD for HT care and HT control, magnitude of association not reported in paper. | Unclear risk of sample bias. |
| Moy et al. 1995 [55] | US. Nationally representative samplen = 6,158 | Cross-sectional | OR for no treatment (reference 1 for general or family practitioner), Internist OR = 0.82 (0.67–1.00), Non primary care physician OR = 1.20 (0.97–1.49), No particular physician OR = 2.61 (2.15–3.18) | High risk of non-differential misclassification. Unclear risk of differential misclassification. |
| Shea et al. 1992b [60] | US. Hospital-based African American and Hispanic inner-city populationn = 207 | Cross-sectional | OR for non-adherence for lack of primary care physician 2.9 (1.36–6.02 versus presence of primary care physician. | High risk of sample bias. Unclear risk of non-differential misclassification bias. |
| Victor et al. 2008 [72] | US. Mostly non-Hispanic African Americans from Dallas Countyn = 1514 | Cross-sectional | OR for HT awareness 3.81 (2.86–5.07), treatment 8.36 (5.95–11.74), and control 5.23 (3.30–8.29): Has a regular physician versus has no regular physician. | Low risk of bias. |
| Routine physician or place of care for HT | ||||
| Ahluwalia et al. 1997 [36] | US. Low-income, African-Americans in an urban ambulatory hospitaln = 221 | Case-control | OR of HT control: Regular source of care 7.93 (3.86–16.29) versus no regular source of care. | Unclear risk of differential misclassification. |
| Private versus public provision of care | ||||
| Dennison et al. 2007 [44] | South Africa. Peri-urban black South Africansn = 403 | Cross-sectional | No significant effect of provider type on systolic BP or odds of BP control below threshold (>140 mmHg systolic and >90 mmHg diastolic BP). Diastolic BP 3.29 mmHg greater in public versus private sector (p = 0.042). | Unclear risk of sample bias. |
| Kotchen et al. 1998 [75] | US. Inner-city African American population from Milwaukeen = 583 | Cross-sectional | Unadjusted OR for HT control: Private provider 1.20 (0.62–2.32) versus non-private provider | High risk of confounding. Unclear risk of sample bias. |
| de Santa-Helena et al. 2010 [42] | Brazil. Patients from family health units in Blumenaun = 595 | Cross-sectional | OR for non-adherence: Treated by public health service (SUS) 1.8 (1.1–2.7) versus private medical provider. | Unclear risk of non-differential misclassification. |
| Yiannakopoulou et al. 2005 [74] | Greece. Patients admitted for elective surgery in Athens.n = 1,000 | Cross-sectional | Medication adherence with private physician 25.1% versus 10% of those with physician in rural areas and 8.8% of with physician from the National Health System (p<0.005 between groups) | High risk of confounding. Unclear risk of non-differential misclassification. |
APR, adjusted prevalence ratios.