Abstract
Objectives. This study ascertained the separate and combined effects of having insurance and a usual source of care on receiving preventive services.
Methods. Descriptive and multivariate analyses of 1996 Medical Expenditure Panel Survey data were conducted.
Results. Receipt of preventive services was strongly associated with insurance and a usual source of care. Significant differences were found between insured adults with a usual source of care, who were most likely to have received services, compared with uninsured adults without regular care, who were least likely to have received services. Those with either a usual source of care or insurance had intermediate levels of preventive services.
Conclusions. Having a usual source of care and health insurance are both important to achieving national prevention goals.
Obtaining preventive care services in the US health care system is not an easy task. In the maze of services, providers, and financing arrangements, health insurance coverage has been shown to contribute to better health outcomes.1–3 Despite this evidence, repeated attempts to extend health insurance coverage to everyone in the United States have failed, and recent estimates suggest that nearly 43 million people (16% of the population) have no health insurance.4,5 Although reports indicate that this number might be declining, it has increased from 1989 reports of 33.4 million.4,6 One example of an increase in this population is the growing number of uninsured people in New York State, where 1 in 5 were uninsured in 1997, an increase of 46% from 1990.7
Amid stymied efforts to extend insurance coverage, other strategies for improving access and health outcomes are being explored. For example, a surprisingly large number of uninsured people have been able to establish a regular, ongoing relationship with either a health service facility or an individual provider as a “usual source of care.”8,9 Once people secure a usual source of care, they are more likely than those without a usual source of care to gain access to services, including preventive care8,10–15 and regular physician visits.12 A consistent source of care also has been associated with lower use of the emergency department and shorter lengths of stay in the hospital.16–20 The development of a continuous relationship with a medical caregiver is especially important for people with chronic illnesses and mental health problems.21–23 When people lack a usual source of care, their access to necessary services is reduced,11,12,24–28 which may result in poorer health outcomes.29 When directly compared with health insurance status, usual source of care has been found to be a stronger predictor of access to care30 and to have a stronger influence on the receipt of preventive services10 and the likelihood of timely visits to health care facilities.30
Noting these favorable effects of having a usual source of care, some have concluded that linking every person to a regular provider may be more feasible than pursuing further attempts to implement national health insurance. Some states have outlined explicit policies to ensure that certain groups of people have a regular source of care.13,18,19 The Bush administration has proposed increasing federal support for community health centers and other critical facilities that serve as regular sites of care for millions.31–33 The effects of these strategies depend, in part, on the interactive effects of having health insurance and a usual source of care, and this interaction has not received much scrutiny.
We used nationally representative data to describe the relation of health insurance status and having a usual source of care to receipt of preventive services; people were characterized as having neither, both, or one or the other. We then assessed the extent to which usual source of care and health insurance status influence, together and independently, the likelihood that a person will receive preventive care and thus may represent discrete, important policy options for improving health promotion and disease prevention.
Data used in this cross-sectional study were taken from the 1996 Medical Expenditure Panel Survey data, which were sponsored and made available to the public by the Agency for Health Care Research and Quality.27 The Medical Expenditure Panel Survey consists of information on a nationally representative sample of civilian, noninstitutionalized persons in the United States.34 A sample of households from the previous year’s National Health Interview Survey was designated eligible to participate in the 1996 Medical Expenditure Panel Survey.27 Computer-assisted personal interviewing was used for data collection. Our study was restricted to the 14 995 Medical Expenditure Panel Survey participants aged 18 years or older.
This study included reports from adults (≥18 years of age) about preventive services received during 1996. A total of 9 preventive services were analyzed (5 services among all respondents, 3 services among women only, 1 service among men only). Specific services included blood pressure check, cholesterol check, physical examination, flu shot, dental checkup, Papanicolaou test (women only), breast examination (women only), mammogram (women only), and prostate examination (men only). These particular items were selected because of their inclusion in the Medical Expenditure Panel Survey and because of recent recommendations by the US Preventive Services Task Force about the usefulness of these services in improving health outcomes.35
Data were analyzed in different subsets of the survey population for each of the 9 preventive services based on the US Preventive Services Task Force recommendations to guide inclusion criteria by age and sex. Physical examinations and dental checkups are recommended for adults of all ages; see Table 1 for definitions of subgroups analyzed for each of the other preventive services. Flu shots and prostate examinations were excluded from Table 1 because of the small number of people without both insurance and a usual source of care.
The potentially explanatory variables of primary interest were health insurance coverage and usual source of care. Any insurance coverage, without regard to adequacy, was accepted as having insurance. To determine usual source of care, respondents were asked, “Is there a particular doctor’s office, clinic, health center, or other place that you go to if you are sick or need advice about your health?” For comparative analyses, responses about usual source of care and health insurance were divided into 4 categories: (1) yes usual source of care/yes insurance, (2) yes usual source of care/no insurance, (3) no usual source of care/yes insurance, and (4) no usual source of care/no insurance. Receipt of preventive services is, of course, also influenced by other factors. We controlled statistically for additional variables in an effort to assess the specific effects of insurance status and usual source of care. Demographic characteristics included in the analyses for this purpose were age, sex, race/ethnicity, completion of high school (head of household), residence within or outside a metropolitan statistical area, and perceived health status.
Four categories were constructed to account for all Medical Expenditure Panel Survey participants in terms of their insurance status and whether they had a usual source of care. Simple comparisons were conducted to determine the relation of demographic characteristics to each of the 4 usual source of care and insurance categories (Tables 2 and 3). Multivariate logistic regression was performed to assess associations of usual source of care and insurance with the use of preventive services among age- and sexappropriate subgroups, controlling for potentially confounding factors (Table 1). SUDAAN (Research Triangle Institute, Research Triangle Park, NC) software was used to conduct statistical tests and to make national estimates with the variance adjustment required by the complex sampling design of the 1996 Medical Expenditure Panel Survey. In all tables provided, the number of respondents (n) represents the unweighted survey sample size, whereas the reported percentages have been weighted to produce estimates for the entire US population.
More than 79% of the adults had a usual source of care in 1996 (Table 2). Similarly, nearly 83% had health insurance. Almost 70% had both a usual source of care and insurance, whereas fewer than 8% had neither. More than half of the uninsured adults had a usual source of care (1573 of 2886; 54.5%).
Several demographic variables were strongly related to having health insurance and a usual source of care (Table 3). For example, more than 90% of respondents older than 64 years reported having both a usual source of care and health insurance compared with only 51% of respondents aged between 18 and 24 years. In contrast, nearly 15% of 18- to 24-year-olds reported no usual source of care and no insurance, compared with fewer than 1% in respondents older than 64. A higher percentage of women (74%) were insured and had a usual source of care compared with men (65%). Eleven percent of the Black population and almost 23% of the Hispanic population reported no insurance and no usual source of care, compared with fewer than 7% of Whites. Adults who had completed high school were more likely to be insured and to have a usual source of care (72%) than were those who had not completed high school (61%). Almost the same proportions of people living inside or outside metropolitan statistical areas had both insurance and a usual source of care (70% and 68%, respectively). Fewer than 3% of the adults who perceived their health status as poor were without both insurance and a usual source of care compared with nearly 8% of those who thought themselves to be in excellent health. Among the insured, about 17% of those who believed themselves to be in excellent health had no usual source of care, whereas fewer than 6% of those reporting poor health had no usual source of care.
Receipt of preventive services supported by the US Preventive Services Task Force was strongly associated with being insured and having a usual source of care (Table 1). Flu shots and prostate examinations were not included because of small numbers in some cells. The group of adults with neither a usual source of care nor health insurance was used as the reference group (odds ratio [OR] = 1.0). People with insurance and a usual source of care were the most likely to have received services within the most recent 12 months. A large percentage of people with both a usual source of care and insurance had received 1 or more of the following preventive services within the past 12 months: blood pressure checks, cholesterol checks, physical examinations, dental checkups, Papanicolaou tests, breast examinations, or mammograms. Uninsured people without a usual source of care were the least likely to have received preventive services; for all 7 preventive services, fewer than half of the adults in this subgroup had received the services in the past 12 months.
A blood pressure check (in 46% of adults) was the most common preventive service received by uninsured adults without a usual source of care. Among the women in this subgroup, 35% had received a Papanicolaou test, 27% had received a breast examination, and 16% had received a mammogram. In comparison, 67% of the women with both insurance and a usual source of care had received a Papanicolaou test, 71% had received a breast examination, and 57% had received a mammogram. Only 16% of the uninsured adults without a usual source of care had had their cholesterol levels checked compared with 59% of the adults with both a usual source of care and insurance. About 54% of the adults with both a usual source of care and insurance had received a physical examination, whereas only 23% of the subgroup without both a usual source of care and insurance had received this service. Similarly, dental checkups were received by 41% of the subgroup with both a usual source of care and insurance compared with only 13% of the uninsured without a usual source of care.
For breast examinations and mammography, the confidence intervals for odds ratios for subgroups lacking either insurance or a usual source of care contained 1.0, as did 1 of the subgroups for blood pressure checks, cholesterol checks, and Papanicolaou tests. A consistent pattern was found, with likelihood of preventive services being highest for those with both insurance and a usual source of care, lowest for those with neither, and intermediate for those with one or the other.
In 1991, the US government published Healthy People 2000: National Health Promotion and Disease Prevention Objectives with a goal “to improve the financing and delivery of clinical preventive services so that virtually no American has a financial barrier to receiving at a minimum screening, counselling and immunization services.”3636(p215) As shown in Table 1, after control for several demographic variables, insurance and a usual source of care had independent, additive effects on the receipt of preventive services.
Our study found statistically significant benefit in having both a usual source of care and insurance, which was the optimal condition relative to being uninsured and without a usual source of care. The results were mixed when we compared the 2 “halfway” groups (yes usual source of care/no insurance, no usual source of care/yes insurance). In contrast to findings of some recent studies, having a usual source of care was not always a more consistent predictor of the receipt of preventive services when we compared people who had either insurance or a usual source of care but not both. Uninsured Americans without a usual source of care are unlikely to receive adequate preventive services. Thus, ensuring that all Americans have both health insurance and a usual source of care is the best way to reach the goals for clinical preventive services set forth in Healthy People 2000.
What can be done to help uninsured Americans without a usual source of care, who have the greatest risk of not receiving adequate preventive services? This study showed that access to health insurance and a usual source of care is not randomly distributed among the US population. Efforts aimed at increasing the number of people who have a usual source of care and insurance should target Hispanic and non-White subgroups and those living in households headed by individuals lacking a high school education.
Although emphasizing primary care is a logical strategy to improve preventive services, it is prudent to be wary of becoming overly reliant on a safety net known to be fragile.7,31,37–39 The value of continuous primary care is well established.10,26,40–42 Yet that value is dependent on a system relatively free of disruptions of care. Threats to continuity of care, on which the usual-source-of-care relationship depends, undermine the preventive efforts of providers37,43,44 and such threats have become a major source of physician’s primary care dissatisfaction.45 Potential reductions in charity care from physicians being financially squeezed by today’s health care market also indicate the need for caution in assessing the benefits that would likely accrue from policies that ensure a usual source of care.31,46–48 Additionally, as noted in this study and elsewhere, access to a usual source of care is not a guarantor of prevention services, nor does it ensure availability of prescriptions, specialty care, certain needed procedures, or home care services.7
This study adds further reasons to be concerned about increasing strains on US safety-net facilities, the rising number of uninsured people, and inequities in the accessibility of services. Improving preventive service delivery to the entire US population requires expanding health insurance coverage and improving access to comprehensive and continuous primary care services. Our data also indicate that even under the best of circumstances, there is room for improvement in the delivery of preventive services. For example, in the group of adults with both a usual source of care and health insurance, fewer than half had received dental services, and only 54% had received the physical examination recommended by the US Preventive Task Force as a way to provide an opportunity for early detection of cancers and other illnesses.
As important as health insurance and a usual source of care are to receipt of preventive services, they do not in themselves ensure adequate access. In 1996, Medical Expenditure Panel Survey respondents reported difficulty or delay in obtaining needed health care owing to transportation or communication problems as well as to their own physical problems. Others did not have time, child care, or authorization to miss work.27 Certain of the reasons for not having a usual source of care were related to health insurance. Respondents cited changing health plans, the cost of insurance, and not having a provider in their plan available nearby as reasons for not having a usual source of care.49
This analysis had important limitations. As in all surveys, responses are subject to possible reporting error and to response biases not accounted for by statistical adjustments. Our findings are associations between variables and do not establish causal relationships. Uncertainties remain regarding how to define “having health insurance” and “having a usual source of care.” Neither the dollar amount nor the services or settings of care covered by insurance were specified. The usual source of care could be a facility or an individual health professional. The Medical Expenditure Panel Survey does not permit determination of the proportion of care a respondent received that was provided by a usual source of care.
Although having a usual source of care was not consistently shown to be superior to having health insurance, access to a usual source of care may be more achievable through local initiatives and through some less costly approach than guaranteeing universal health care coverage to expanding the delivery of prevention services. However, neither approach displaces the need for the other. Having a usual source of care and having health insurance are independent and additive predictors of the likelihood of receiving preventive care. Both should be pursued to meet our nation’s goals for preventive service delivery.
aAged 21 years or older. bWomen aged 45–64 and men aged 35–64 years. cWomen aged 18–64 years. dWomen aged 40–69 years. Note. MSA = metropolitan statistical area. aNs/ns are unweighted, and the percentage of this nationally representative sample is weighted.
Procedure n (Unweighted) Percentage Receiving Preventive Services (Weighted) Multivariate Odds Ratio (95% Confidence Interval) Blood pressure check Totala 13 827 78 Yes usual source of care/yes insurance 9475 85 4.83 (2.40, 9.70) Yes usual source of care/no insurance 1387 70 4.43 (2.12, 9.25) No usual source of care/yes insurance 1830 62 1.90 (0.82, 4.39) No usual source of care/no insurance 1135 46 1.00 Cholesterol check Totalb 5602 51 Yes usual source of care/yes insurance 3944 59 4.51 (1.50, 13.56) Yes usual source of care/no insurance 574 45 4.61 (1.28, 16.61) No usual source of care/yes insurance 701 30 1.67 (0.38, 7.36) No usual source of care/no insurance 383 16 1.00 Physical examination Total 14 536 47 Yes usual source of care/yes insurance 9840 54 3.17 (2.59, 3.87) Yes usual source of care/no insurance 1528 38 1.98 (1.58, 2.48) No usual source of care/yes insurance 1925 30 1.35 (1.09, 1.67) No usual source of care/no insurance 1243 23 1.00 Dental checkup Total 14 825 35 Yes usual source of care/yes insurance 10 011 41 3.50 (2.75, 4.46) Yes usual source of care/no insurance 1556 19 1.43 (1.10, 1.87) No usual source of care/yes insurance 1972 30 2.26 (1.76, 2.89) No usual source of care/no insurance 1286 13 1.00 Papanicolaou test Totalc 6222 62 Yes usual source of care/yes insurance 4220 67 7.27 (2.76, 19.13) Yes usual source of care/no insurance 829 49 1.95 (0.59, 6.44) No usual source of care/yes insurance 707 56 4.13 (1.22, 14.01) No usual source of care/no insurance 466 35 1.00 Breast examination Totald 3413 65 Yes usual source of care/yes insurance 2591 71 14.13 (1.76, 113.63) Yes usual source of care/no insurance 356 47 3.44 (0.38, 31.00) No usual source of care/yes insurance 300 49 7.74 (0.64, 93.18) No usual source of care/no insurance 166 27 1.00 Mammogram Totald 3403 52 Yes usual source of care/yes insurance 2578 57 17.53 (1.49, 206.60) Yes usual source of care/no insurance 358 36 3.89 (0.28, 54.56) No usual source of care/yes insurance 299 38 3.29 (0.18, 59.13) No usual source of care/no insurance 168 16 1.00 n (Unweighted) Percentage Receiving Preventive Services (Weighted) Yes usual source of care 11 689 79.2 No usual source of care 3306 20.8 Yes insurance 12 109 82.8 No insurance 2886 17.2 Yes usual source of care/yes insurance 10 116 69.5 Yes usual source of care/no insurance 1573 9.7 No usual source of care/yes insurance 1993 13.3 No usual source of care/no insurance 1313 7.5 n (% of Total) Yes Usual Source of Care/Yes Insurance, n (%)a Yes Usual Source of Care/No Insurance, n (%)a No Usual Source of Care/Yes Insurance, n (%)a No Usual Source of Care/No Insurance, n (%)a Total 14 995 10 116 (69.48) 1573 (9.72) 1993 (13.29) 1313 (7.50) Age in 1996, y (N = 14 995) 18–24 1833 (12.78) 880 (51.10) 329 (17.76) 298 (16.55) 326 (14.59) 25–44 6396 (42.88) 3882 (63.25) 765 (11.02) 1018 (15.91) 731 (9.82) 45–64 4323 (27.73) 3165 (75.24) 455 (9.42) 465 (10.64) 238 (4.71) ≥ 65 2443 (16.61) 2189 (90.10) 24 (0.67) 212 (8.52) 18 (0.71) Sex (N = 14 995) Female 8050 (52.18) 5766 (73.91) 897 (9.91) 866 (10.96) 521 (5.22) Male 6945 (47.82) 4350 (64.64) 676 (9.51) 1127 (15.86) 792 (9.99) Race (N = 14 995) American Indian 181 (1.11) 105 (58.66) 33 (15.93) 24 (15.23) 19 (10.18) Aleut, Eskimo 8 (0.07) 6 (79.91) 2 (20.09) 0 (0.00) 0 (0.00) Asian or Pacific Islander 440 (3.53) 279 (63.34) 55 (12.50) 55 (12.31) 51 (11.85) Black 1941 (11.67) 1251 (61.22) 250 (14.95) 249 (12.83) 191 (11.00) White 12 410 (83.58) 8472 (71.05) 1231 (8.78) 1662 (13.39) 1045 (6.79) Other 15 (0.04) 3 (28.29) 2 (7.38) 3 (31.35) 7 (32.98) Ethnicity (N = 14 995) Hispanic 2711 (9.73) 1244 (46.41) 442 (15.97) 416 (14.98) 609 (22.65) Black/not Hispanic 1844 (11.25) 1193 (61.50) 240 (15.00) 238 (12.78) 173 (10.72) Other 10 440 (79.02) 7679 (73.46) 891 (8.20) 1339 (13.17) 531 (5.18) Completed high school (N = 14 995) Yes 11 388 (80.06) 8056 (71.58) 1037 (8.65) 1552 (13.65) 743 (6.11) No 3607 (19.94) 2060 (61.02) 536 (14.01) 441 (11.90) 570 (13.07) Urban/rural (N = 14 995) MSA 11 726 (80.38) 7921 (69.77) 1111 (8.76) 1609 (13.65) 1085 (7.82) Non-MSA 3269 (19.62) 2195 (68.28) 462 (13.64) 384 (11.86) 228 (6.21) Perceived health status (N = 14 988) Excellent 4140 (28.85) 2671 (66.10) 381 (8.97) 702 (17.04) 386 (7.89) Very good 4793 (32.95) 3272 (70.07) 444 (8.58) 670 (14.23) 407 (7.12) Good 3931 (25.27) 2592 (68.73) 485 (11.08) 471 (11.45) 383 (8.74) Fair 1557 (9.40) 1118 (75.16) 204 (12.21) 118 (6.55) 117 (6.08) Poor 567 (3.53) 459 (81.84) 59 (10.14) 31 (5.36) 18 (2.67)
J. E. DeVoe was the principal author and conceptualized the research with significant input from G. E. Fryer, R. Phillips, and L. Green. G. E. Fryer conducted the analyses. All of the authors participated in the review and revisions on drafts of the article.
Human Participant Protection No protocol approval was needed for this study.
