Impact of Work Conditions and Minority Patient Populations on Quality and Errors

Abstract

Objectives:

To determine whether workplace conditions affect care quality and errors, especially in primary care clinics serving minority patients.

Methods:

We conducted a 3-year assessment of work conditions and patient outcomes in 73 primary care clinics in the upper Midwest and New York City. Study participants included 287 physicians and 1204 patients with hypertension and/or diabetes. Chart audit data were contrasted between clinics with ≥30% minority patients (minority-serving clinics, or MSCs) and those with <30% (nonminority-serving clinics, or NMSCs). Physicians reported on time pressure, work control, clinical resources, and specialty referral access; managers described room availability; and chart audits determined care errors and quality. Two-level hierarchical models tested work conditions as mediators between MSC status and clinical outcomes.

Results:

Error rates were higher in MSCs than NMSCs (29.6% vs 24.8%, P < .05). Lack of clinical resources explained 41% of the effect of MSC status on errors (P < .05). Diabetes control was poorer in MSCs than in NMSCs (53.8% controlled vs 76.1%, P < .05); lack of clinical resources explained 24% of this difference (P < .05). Room availability increased quality in both MSCs and NMSCs by 5.95% for each additional room per clinician per session. Lack of access to rooms and specialists decreased the likelihood of blood pressure control in MSCs.

Conclusion:

Work conditions such as clinical resources, examination room availability, and access to referrals are significantly associated with errors and quality, especially in MSCs.

Introduction

Racial disparities in health care may be influenced by the sites where patients receive care.^1-2 One study showed that black patients were more likely to attend hospitals with lower rates of evidence-based medical treatments,² while another showed that risk-adjusted mortality after acute myocardial infarction was greater in hospitals with high proportions of black patients.³ Health disparities research has recently focused on addressing these decrements in quality.

The Minimizing Error Maximizing Outcome (MEMO) Study found more challenging work conditions in primary care clinics that served minority patients.⁴ Physicians practicing in these clinics had less access to specialists, clinical resources, and examination rooms. Doctors in these clinics reported more chaotic workplaces, less control, and lower job satisfaction. Based on these troubling findings, we examined data from MEMO to answer the following questions: Are there differences in quality of care or errors between clinics serving large proportions of minority patients and those serving mainly nonminority patients, and, if so, how far do work conditions explain these differences?

Method

The 2002-2005 MEMO study was a longitudinal investigation of primary care clinics in New York city; Chicago, Illinois; and rural, academic, and urban settings in Wisconsin.⁵ Eligible participants were family practitioners or general internists with at least 4 sessions per week in ambulatory care. Up to 6 patients per physician were recruited by mail or in person. Eligibility criteria included age 18 or older; ability to read English, Spanish, or Cantonese; at least 2 outpatient visits with a study physician in the previous year; and a diagnosis of hypertension, diabetes mellitus, and/or congestive heart failure. (Heart failure patients were eventually eliminated from analysis due to systematic challenges in data collection.) Clinics serving ≥30% minority populations based on manager data were deemed minority-serving clinics (MSCs). This threshold resembles the 2005 U.S. proportion of Blacks plus Hispanics.⁶

A physician survey queried satisfaction, stress, burnout, work control, referral access, office pace (calm to chaotic), access to clinical resources, and likelihood to leave within 2 years. Managers reported on examination room availability. Patient records were audited retrospectively over 18 months for errors, quality, diabetes control, and hypertension control based on national guidelines. Error points were assigned for missing processes of care, including depression screening, diet/exercise advice, tobacco use documentation, and cancer screening. Quality points were assigned for control of diabetes (Hemoglobin A1c ≤7.5%) and hypertension (blood pressure <140/90 or <130/85 for diabetic patients) for ≥50% of recorded measurements.⁷

Two-level hierarchical patient–physician structural models tested the following: (1) associations between MSC status and errors, quality, diabetes control, and hypertension control and (2) the mediating effects of time pressure, chaos, work control, resources, referral access, and examination rooms. Models were adjusted for patient and physician demographics. Figure 1 illustrates the analysis’ conceptual basis.

Figure 1.

Conceptual model.

Results

The MEMO recruited 119 clinics, 422 physicians (54% of those approached), and 1795 patients. Seventy-three clinics had complete data (287 physicians and 1204 patients). Among physicians, 63% were general internists and 48% were females. Among patients, 65% were females, 54% had diabetes, 90% had hypertension, and 44% had both. Of the 73 clinics, 36% were MSCs. Parameter estimates for the models are available from the authors upon request.

Error Models

The MSC status was associated with more time pressure and chaos; less room availability, work control, and referral access; and fewer resources. Adjusted error rates were almost 5% higher in MSCs (29.6% vs 24.8%, P < .05). Fewer clinical resources explained 41% of this effect.

Quality models

Each additional examination room per physician increased quality scores by 5.95%. Of note, there were fewer rooms in MSCs (2.2 per clinician vs 2.7 in non-MSCs [NMSCs], P < .05).

Hypertension Control Models

Blood pressure control was similar between MSCs and NMSCs. Referral access and number of examination rooms (both P < .05) partially explained blood pressure control in MSC patients.

Diabetes Control Models

Diabetes control was poorer in MSCs (53.8% controlled vs 76.1%, P < .05). Resource access partially mediated this effect, explaining 24% of the difference.

Discussion

Our previous work has uncovered challenging work conditions in MSCs.⁴ In this study of 287 primary care physicians and 1204 patients, we found greater overall errors and poorer diabetes control in clinics serving larger proportions of minority patients. The study provides evidence that one of these conditions, access to clinical resources, explains in large part the 5% difference in overall errors between MSCs and NMSCs and partially accounts for better diabetes control in NMSCs. Our data also provide evidence that examination room availability and access to specialists are associated with improved control of blood pressure and higher overall quality.

One key finding from this study is the importance of sufficient examination rooms. Fewer examination rooms were found per provider per session in MSCs. Diminished room availability was linked to lower quality of care. We thus encourage all clinics, especially those serving minority patients, to allow sufficient space for conducting clinical examinations and providing high-quality care.

Limitations of our study include self-reported physician data and patients residing in few geographical regions. While data are from 2002 to 2005, they remain pertinent due to the strong study design and likely stability of relationships between predictor variables and clinical outcomes. The MEMO’s many strengths (including a diverse patient population, a large number of clinics, and objective outcome data) support the importance of the findings.

Author Biographies

Anita B. Varkey is a general internist at Loyola University Medical Center. She is the Associate Director of Primary Care and is also the Medical Director of the General Medicine Clinic at Loyola’s Outpatient Center. Her clinical, educational and research areas focus on ambulatory care of chronic medical diseases and prevention.

Linda Baier Manwell is the Research Program Administrator for the Division of General Internal Medicine at the University of Wisconsin School of Medicine and Public Health. Since 2007, she has also been the National Training Coordinator for Women’s Health Education at VA Women’s Health Services. She has a long history of developing and implementing research and education projects focused on alcohol abuse, physician burnout, and women’s health.

Roger L. Brown is a Professor and Director of the Research Design and Statistics Unit, University of Wisconsin-Madison, School of Nursing, Medicine and Public Health.

Enid Montague received MS and PhD degrees in Industrial and Systems Engineering from Virginia Tech in 2008, specializing in human factors and ergonomics engineering and human computer interaction. Dr. Montague is an Associate Professor in the college of computing at DePaul University and adjunct professor at Northwestern University’s Feinberg School of Medicine. She is director of the Wellness and Health Enhancement Engineering Laboratory (WHEEL). Her research uses human factors and human-computer interaction methodologies, design principles and theories to understand health care systems to promote safety and patient-centered care.

Neda Laiteerapong is a clinical investigator and internist at the University of Chicago. She is committed to improving outcomes for patients with Type 2 diabetes by ensuring that they receive personalized and timely care that addresses both physical and mental health. Her work focuses on tailoring individualized care for adults with diabetes to their life course, prioritization of care for patients with type 2 diabetes, the intersection of health policy and the care of patients with chronic diseases, and the impact of chronic diseases on quality of life.

Diana Burgess received her Ph.D. in Psychology from the University of Minnesota in 1998. She is currently Associate Professor of Medicine at the University of Minnesota and a Principal Investigator and Co-Director of the Veterans Affairs Advanced Fellowship Program in Health Services Research at the Center for Chronic Disease Outcomes Research, Minneapolis Veterans Affairs Healthcare System, Minneapolis, MN.

Said Ibrahim is the Co-Director of the VA HSR&D Center of Innovation for Health Equity Research and Promotion (CHERP), and former Associate Chief of Staff for Medicine at Philadelphia VA Medical Center. He is also a tenured Professor of Medicine at the University of Pennsylvania, Perelman School of Medicine.