Black/white differential use of health services by young children in a rural Mississippi community.

Background and Objectives: Use of health services in rural communities has not been well studied. We explored how black and white children used health services in a rural Mississippi community.

Methods: Data were prospectively collected for 396 children attending a private practice to determine if race was associated with the use of health services in this community.

Results: White children made more sick contacts than black children (P < 0.001). Black children (36%) were more likely to be treated in the emergency room than white children (24%; P = 0.013). There was no black-white difference in the hospitalization rates, although white children were more likely to undergo ear-nose-throat (ENT) surgery for pressure equalizing tube (PET) placement, and/or tonsillectomy and adenoidectomy (T & A; P < 0.001). Even after controlling for various confounders, the frequency of all physician sick contacts was twice as high for white children than for black children (RR = 2.17; 95% C.I. = 1.32-3.58).

Conclusion: Overall, black children used disproportionately fewer health services than their white counterparts, but used significantly higher emergency room services. Private insurance coverage was the single most significant variable that accounted for the black-white differential use of the emergency room.

Key Words: use of health services, race/ethnicity, rural health, physician sick contacts

**********

Ethnic and racial disparity in the use of health services has been well documented in the United States. Ethnic minorities tend to use health services less frequently than whites and the quality of services has been shown to be better for whites than for other ethnic minorities. (1-4) The use of health services has been shown to vary by state, ethnicity, maternal education, marital status, insurance status, and perceived health status. (2,5-9) Bartman et al (4) showed that black adults reported fewer physician contacts than their white counterparts in the previous year. Gornick et al (1) found that blacks had lower use of ambulatory and preventive care services than whites even after controlling for demographics, including income level. Even though whites have higher use of health services than blacks and Hispanics, the type of health services used also has been shown to vary by ethnicity. For instance, white women consistently have a higher frequency of prenatal care visits than black or Hispanic women. (1,10-12) Abrams and Nathan (2) showed that, as compared with blacks, whites were more likely to receive certain hospital procedures such as catheterization, bypass surgery, endoscopy, bronchoscopy, cesarean section or organ transplantation. On the other hand, blacks were more likely than their white counterparts to have organ removal (bilateral orchiectomy, limb amputation) after controlling for various confounders. (1) Many national population-based studies conducted predominantly in large urban communities have shown that blacks utilize the emergency room (ER) more often than whites. (5,12,13) Few studies, however, have been done in smaller rural communities to determine if such ethnic disparities also exist.

Because most of rural America is predominantly white and middle income, it is logical to assume that the use of health services would be more equitable than in urban areas. However, the north Delta region of Mississippi is unique in that the population is predominantly impoverished rural black (65% black versus 35% white). It was thought that the black-white differential use of health services would also reflect that of urban America. It was therefore hypothesized that in this small and impoverished rural southern community, as compared with their white counterparts, black children would be: (1) more likely to use the ER, (2) less likely to use physician offices, and (3) more likely to be hospitalized for various illnesses.

Materials and Methods

Place and Time of the Study

The Northwest Mississippi Regional Medical Center (NWMRMC) medical staff that served as the institutional review board at the time that this study was conducted approved the study. Data were collected over a 12-month period from May of 1996 to April of 1997. The place of study was the Children's Clinic of Clarksdale, in Coahoma County, Mississippi. At the time of the study, three board-certified pediatricians staffed the clinic. These physicians were the only pediatricians in the service area. They were also the attending pediatricians at the Northwest Mississippi Regional Medical Center (NWMRMC), the only referral hospital for the other three surrounding counties of Tunica, Tallahatchie, and Quit-man. The clinic also saw patients from surrounding counties which did not have a pediatrician. These counties are located in the rural Delta region of northwest Mississippi. Clarksdale, the largest town in the area, has a population of approximately 20,000 inhabitants, 65% of whom are impoverished blacks with an average annual income of around $10,000.00. The hospital medical staff/ethics committee approved the study.

Design, Sampling, and Eligibility Criteria

A convenience sample of 396 patients (8% of the total number of approximately 5,000 regular clinic patients) was included in the survey. These patients were determined by the recruiting pediatrician to be regularly using our clinic. A regular patient was defined as one who had used the clinic at least twice within the previous year or whose parent stated that this was the only clinic regularly used by their child. All patients were recruited on random days chosen by the clinic staff starting from May 1, 1996 to April 30, 1997. When a patient was initially recruited, various sociodemographic data known from the literature to be associated with health services use were collected through direct interviews of the parents or guardians. Variables included maternal age, educational achievement, marital status, employment, number of people in household, the presence of any pets at home, and type of heating. The smoking status of all family members (parents, grandparents, uncles, aunts or friends) in regular contact with the child was determined through direct questioning of the parents or guardians. Information on infant characteristics such as birth weight, gestational age, sex, infant feeding, day care attendance, family history of asthma or other allergies was further obtained. On each sick visit the clinical diagnosis was duly noted in a logbook. Reasons for clinic visits were recorded and included otitis media, upper respiratory infections (URIs), gastroenteritis, wheezing and or/asthma symptoms, pneumonia or bronchitis, sinus infections, skin infections, skin rashes, pharyngitis, and conjunctivitis. Reasons for ER visits and hospital admissions were also carefully recorded. It was easy to determine the number of ER visits and hospital admissions for each patient since this is the only hospital in the area used by our patients. Because we were also the admitting physicians for our patients, copies of all ER visit reports and hospital admissions were available to us the following morning during hospital rounds. This enabled us to accurately track all ER visits and hospital admissions on a daily basis during the study period.

Definition of Terms

A sick visit or sick contact was defined as a patient seeking care from a physician in the clinic or office, ER, or when hospitalized. When a patient was seen for well child-care, but also was diagnosed and treated for an illness, this was recorded as a sick visit. Well-child visits and immunizations were not recorded as sick visits. Parents identified themselves as either black or white. Other races were not included in the analysis since they were very few in number. In this study, upper and lower respiratory tract infections (RTIs) were defined as pharyngitis, conjunctivitis, viral illness, bronchitis, sinusitis, and pneumonia. A patient was considered to have wheezing/reactive airways disease (RAD) and/or possible asthma if symptoms were completely reversible by the use of bronchodilators. A child with three or more of these episodes within the course of a year was diagnosed with probable asthma. No spirometry after methacholine challenge was done to validate the diagnosis of asthma. Acute otitis media (AOM) or acute otitis media with effusion (AOME) was diagnosed using the following standard procedures: a patient presenting with symptoms of fever, fussiness, or ear pain and found on pneumatic otoscopy to have (1) a bulging, dull, and/or red ear drum with or without effusion, loss of bony landmarks; and (2) decreased mobility of the ear drum. A child was considered exposed to tobacco smoke if any of the parents or other relatives admitted to smoking, regardless of whether they denied smoking in the presence of the child and regardless of the number of cigarettes smoked per day. A patient was considered to be a regular patient if the parents stated that we were the only physicians who treated their child except for ER visits. History of breastfeeding was defined as any breastfeeding after delivery, regardless of duration. A child was considered to be attending day care if the parents reported attending at least 3 times weekly and the intention was to keep the child in day care for the following year. Patients who had more than 7 visits during the one-year study period were regarded as high frequency users of office visits or contacts.

Data Analysis

The primary independent variable of interest was race. The outcome variable of interest was differential use of health services through office visits, ER visits, and hospital admissions. A secondary outcome variable of interest was type of diagnosis: upper and lower respiratory infections, asthma and/or reactive airway disease, and middle ear disease. Univariate analysis was first done to determine if there were any ethnic differences in the frequency of physician sick contacts for RTI, AOM/AOME, and asthma/RAD. A [chi square] test was used in the analysis to determine differences in proportions for categorical variables with Yates' correction for cell sizes less than 4 and a Student's t test was used for continuous variables when applicable. Multiple logistic regression analysis was then done to determine if there were independent ethnic differences in the use of clinic/office visits, ER visits, or hospitalizations after controlling for the various demographic correlates. Variables used in the multiple logistic regression analysis were appropriately categorized. SPSS (SPSS, Inc., Chicago, IL) and BMDP (Statistical Solutions, Saugus, MA) statistical software were used for these analyses.

Results

Characteristics of the Study Population

Table 1 shows the characteristics of the subjects by race. Black mothers were younger and their children had lower birth weights than white mothers. There was no significant difference in their educational achievement by race. The average household size was higher for black families. Furthermore, black mothers were more likely to be unmarried (P < 0.001) or lack private health insurance (54% for whites and only 3.8% for blacks; P < 0.001) while their white counterparts were more likely to breastfeed, smoke cigarettes, own a pet, and work out of the home. Furthermore, Medicaid covered 73.9% of the sample, while private insurance covered 22.3%. Only 3.5% were without any form of health insurance. This distribution basically reflected the overall nature of insurance coverage of the community when this study was carried out.

Physician Office Sick Contacts

Differences in use of the various health services were explored. In the univariate analysis, white children had many more physician office sick visits than black children (P < 0.001). Also, white children had many more sick contacts for selected diseases such as middle ear disease (P < 0.001), and upper and lower respiratory tract infections (P = 0.01). Although black children had more visits for wheezing/asthma symptoms than white children, the difference was not statistically significant (Table 2). In a multiple logistic regression analysis (see Table 3), white children were again more likely to be treated in the physician offices than their black counterparts even after controlling for various demographics (RR = 2.2; 95% CI = 1.42-3.33). Maternal variables controlled for in the regression analysis included age, education, marital status, and employment status. Infant factors controlled for in the analyses included age of child, birth weight, gender, day care attendance, total number of people living in the house, type of infant feeding, and family history of atopic diseases. Environmental factors were type of heating, exposure to smoking at home (from family members), and exposure to pets at home.

Emergency Room Visits

In the univariate analysis, black children (36%) had a higher frequency of ER visits than their white counterparts (24%; P = 0.013; also see Figure) during the study period. As shown in Table 2, although black children also had a higher frequency of ER visits (5%) than their white counterparts (2.0%) for middle ear diseases, the difference was not significant. Furthermore, as compared with their white peers (16%), black children (27%) were more likely to be treated at least once in the ER for upper and lower respiratory tract infections (P < 0.01). Once again, there were no differences between the two groups in the frequency of asthma symptoms/RAD, although the trend was toward higher use for black children. In a forward stepping multiple regression analysis (Table 3), white children were less likely to use the ER than black children even after controlling for the various demographic variables except for insurance status (RR = 0.51; 95% CI = 0.01-0.81). However, when insurance was entered in the model, the difference in ER visits between black and white children was no longer significant (RR = 0.91; 95% CI = 0.45-1.67). Because of smallness in numbers, it was not possible to study the ethnic differences in the use of the ER for each of the selected illnesses in a multiple logistic regression.

Hospital Admissions

In Table 2, there were no ethnic differences in the total number of hospital admissions for all acute illnesses (12% for whites and 11.0% for blacks). The number of hospitalizations for each of the selected diseases studied, ie, asthma/wheezing symptoms, RTI, and middle ear disease, were too small to produce any meaningful comparisons. However, white children were almost 3 times more likely (P < 0.001) to undergo ear-nose-throat surgery for adenotonsillectomy (including pressure equalizing tube placement; also see Figure).

[FIGURE OMITTED]

Use of All Health Services

Overall, white children had more sick contacts (office visits + ER visits + hospitalizations) than their black counterparts (P = 0.01; also see Table 2). Even after controlling for demographics in a multiple regression analysis, white children used health services twice as often as black children (RR = 2.17; 95% CI = 1.32-3.58).

Discussion

Our study has attempted to show that there is also a disparity in the use of health services between white and black children in this small rural Mississippi community. Population-based studies demonstrating racial disparity in the use of health services among children in the United States are well documented. However, we are not aware of any study done in the rural communities of the United States comparing ethnic differences in the use of health services. Our findings are generally similar to national studies. The findings in the present study are consistent with the findings by Flores et al (13) who demonstrated that minority children in the United States had fewer doctor visits than white children, even after adjusting for various demographics. In the present study, white children used health services twice as often as black children when they were ill, even after controlling for all demographics. Most studies consistently find that racial minority groups tend to use the ER more that their white peers who, on the other hand, are more likely to be seen in physician offices. (14-18) Lieu et al (19) found that minority adolescents used fewer ambulatory physician visits than their white counterparts, despite having worse reported health status. Shi et al (20) also showed that the usual source of care among ethnic minorities was more likely to be in the hospital setting than in physician offices. In many of these studies, ethnicity remained a significant predictor of physician office visits even after controlling for various confounders.

Our study demonstrated that black children in a rural community used the ER one and a half times more often than their white counterparts, in part, confirming our hypothesis. Thus, poor rural blacks face the same difficulties in the use of health services as compared with their urban counterparts. Our study found that black children were diagnosed less frequently in physician offices with middle ear diseases (AOM/AOME), and RTI than their white peers. This finding was rather surprising. Ethnic differences in various illnesses do not imply that blacks are less susceptible. Rather, it could suggest that black children with these diseases may go undiagnosed for longer periods since they use physician offices less frequently. The decreased use of ambulatory services by minority children may explain why the prevalence of middle ear disease found in some national studies is higher in white than in minority children. (21-23) Indeed, Paradise et al (24) demonstrated that when children were routinely examined at frequent intervals for middle ear disease, there were no prevalent ethnic differences.

In the present study, there was a trend toward higher use of ER visits for the selected illnesses among black children than white children, although it was more significant for RTI. This was not surprising, since ethnic minority children tend to use the ER for a variety of illnesses. However, there was no black-white differential in the ER visits for asthma/RAD. This was rather surprising, as most studies in large urban centers have shown a higher prevalence rate of asthma symptoms among black children who are also more likely to be treated in the ER than white children. There was no difference in the rate of asthma symptoms between the two groups of subjects. This lack of significance could be attributed to the small size of the study.

Low income level, lack of private insurance, and poor educational achievement are some reasons often cited to explain away the disparity in use of health services between whites and minority populations. Studies have shown that, even after controlling for insurance and other confounders, black children were still less likely to use preventive health services than their white counterparts. This disparity could suggest that other reasons such as cultural factors and patient-physician incompatibility may also be important in explaining the differential use of health services in America. Thus it is quite possible that more white parents in this rural community buy health insurance because they anticipate more use than their counterparts after controlling for other confounders. Another possible explanation could be the difficulty or the inconvenience that blacks may feel in maneuvering the health system. Our findings are consistent with those of other studies in demonstrating the differential use of health services as found in large urban centers. However, the finding in our study that black children had less physician sick contacts in the office than their white counterparts was even more surprising because accessibility was not an issue. Three pediatricians, two of whom were minorities, managed the private pediatric practice. Yet, white children had higher rates of sick contacts than their black counterparts and were more likely to undergo surgical procedures such as PET placement and T & A even after adjusting for various confounders. Guendelman and Schwalbe (6) suggested that improved insurance coverage for minorities was an important strategy to facilitate access to health care, as private insurance reduces ER visits, and thus, helps reduce overall health expenditures. On the other hand, we also must find ways to encourage the early use of office visits during an illness.

Several limitations in this study need to be addressed. First there is the problem of external validity that was not sufficiently addressed in this study since this was a convenience sample. Second, the study did not assess the use of preventive health services in the sample so it is not possible to compare the immunization rates between the white and black children in this population. As a result, it is possible that differences in the use of health services for sick visits between the two ethnic groups may influence health-seeking behavior. Third, the sample size was small and may explain why there were no differences in the overall hospitalization rates between the ethnic or racial groups. Nonetheless, our findings show a general trend consistent with the findings of other studies.

Conclusion

This study suggests that even in some rural communities, there also exists a black-white differential use of health services. Having private insurance was the single most important variable that accounted for this differential use of ER services, but when this variable was controlled for, the racial difference was no longer significant.

References

1. Gornick ME, Eggers PW, Reilly TW, et al. Effects of race and income on mortality and use of services among Medicare beneficiaries. N Engl J Med 1996;335:791-799.

2. Abrams J, Nathan H. Recipient race does not influence waiting time for a cadaveric renal transplant-one organ procurement organization's experience. Transplant Proc 1991;23:2607-2609.

3. Hargraves JL, Hadley J. The contribution of insurance coverage and community resources to reducing racial/ethnic disparities in access to care. Health Serv Res 2003;38:809-829.

4. Bartman BA. Women's access to appropriate providers within managed care: implications for the quality of primary care. Womens Health Issues 1996;6:45-50.

5. Hargraves JL. The insurance gap and minority health care. 1997-2001. Track Rep 2002;2:1-4.

6. Alexander GR, Cornely DA. Racial disparities in pregnancy outcomes: the role of prenatal care utilization and maternal risk status. Am J Prev Med 1987;3:254-261.

7. Kogan MD, Alexander GR, Kotelchuck M, et al. Relation of the content of prenatal care to the risk of low birth weight. Maternal reports of health behavior advice and initial prenatal care procedures. JAMA 1994;271:1340-1345.

8. Waidmann TA, Rajan S. Race and ethnic disparities in health care access and utilization: an examination of state variation. Med Care Res Rev 2000;57 (Suppl 1):55-84.

9. Reichman NE, Kenney GM. Prenatal care, birth outcomes and newborn hospitalization costs: patterns among Hispanics in New Jersey. Fam Plann Perspect 1998;30:182-187, 200.

10. Weinick RM, Krauss NA. Racial/ethnic differences in children's access to care. Am J Public Health 2000;90:1771-1774.

11. Frisbie WP, Echevarria S, Hummer RA. Prenatal care utilization among non-Hispanic Whites, African Americans, and Mexican Americans. Ma-tern Child Health J 2001;5:21-33.

12. Kogan MD, Kotelchuck M, Alexander GR, et al. Racial disparities in reported prenatal care advice from health care providers. Am J Public Health 1994;84:82-88.

13. McConnel CE, Wilson RW. Racial and ethnic patterns in the utilization of prehospital emergency transport services in the United States. Prehospital Disaster Med 1999;14:232-235.

14. Flores G, Bauchner H, Feinstein AR, et al. The impact of ethnicity, family income, and parental education on children's health and use of health services. Am J Public Health 1999;89:1066-1071.

15. Halfon N, Newacheck PW, Wood DL, et al. Routine emergency department use for sick care by children in the United States. Pediatrics 1996;98:28-34.

16. Cornelius LJ. Barriers to medical care for white, black, and Hispanic American children. J Natl Med Assoc 1993;85:281-288.

17. Walls CA, Rhodes KV, Kennedy JJ. The emergency department as usual source of medical care: estimates from the 1998 National Health Interview Survey. Acad Emerg Med 2002;9:1140-1145.

18. Paradise JL, Rockette HE, Colborn DK, et al. Otitis media in 2253 Pittsburgh-area infants: prevalence and risk factors during the first two years of life. Pediatrics 1997;99:318-333.

19. Lieu TA, Newacheck PW, McManus MA. Race, ethnicity, and access to ambulatory care among US adolescents. Am J Public Health 1993;83: 960-965.

20. Shi L. Experience of primary care by racial and ethnic groups in the United States. Med Care 1999;37:1068-1077.

21. Shurin PA, Pelton SI, Donner A, et al. Persistence of middle-ear effusion after acute otitis media in children. N Engl J Med 1979;300:1121-1123.

22. Hardy AM, Fowler MG. Child care arrangements and repeated ear infections in young children. Am J Public Health 1993;83:1321-1325.

23. Griffith TE. Epidemiology of otitis media--an interracial study. Laryngoscope 1979;89:22-30.

Philip Kum-Nji, MD, MPH, Carole L. Mangrem, MD, Peggy J. Wells, MD, Lisa M. Klesges, PhD, and Henry G. Herrod, MD

From the Children's Clinic of Clarksdale, Clarksdale, MS; the Division of Epidemiology, Department of Health Sciences Research, Mayo Clinic College of Medicine, Rochester, MN; and the Department of Pediatrics, Le Bonheur Children's Medical Center, University of Tennessee College of Medicine, Memphis, TN.

Reprint requests to Philip Kum-Nji, MD, MPH, Children's Medical Center, Virginia Commonwealth University School of Medicine, 1001 E Marshal Street, Richmond, VA 23298; Email: [e-mail]

The authors have no disclosures to declare.

Accepted May 1, 2006.

RELATED ARTICLE: Key Points

* This is the first study to investigate the use of health services in this impoverished rural Mississippi community.

* Overall, black children were less likely to use health services than white children, even controlling for various sociodemographics.

* Insurance status was the single most important variable that explained the black-white differential use of the emergency room.

Table 1. Characteristics of the study population

                              Whites          Blacks
Variables                     (n = 81)        (n = 315)       P value

Continuous variables          Mean (SD)       Mean (SD)
  Maternal age (years)          25.81 (6.15)    23.59 (5.93)    0.006
  Maternal education (years)    12.89 (2.85)    12.27 (2.34)    0.086
  Child's age (months)          21.89 (17.4)    17.89 (15.8)    0.071
  Birth weight (gm)           3351 (617)      3131 (655)        0.008
  Number in household            4.1 (1.05)      4.73 (1.74)  < 0.001
Categorical variables         No. (%)         No. (%)
  Mother smokes                 32 (39.5)       54 (17.1)       0.004
  Male infant                   30 (37.0)       89 (28.3)       0.893
  Day care                      26 (32.1)       53 (16.8)       0.101
  Unmarried mother              18 (22.2)      180 (57.1)     < 0.001
  Breastfeeding (any)           29 (36.0)       41 (13.0)     < 0.001
  Pets at home                  18 (22.2)       22 (7.0)        0.006
  Insurance (private)           44 (54.3)       12 (3.8)      < 0.001
  Type of heating (central)     50 (61.7)      125 (39.7)       0.065
  Working mother (yes)          49 (60.5)       82 (26.0)     < 0.001
  Gestational age < 37          13 (16.0)       45 (19.2)       0.524

Table 2. Use of health services among infants and young children in a
low-income rural Mississippi community

                                    Number of episodes
                                    Whites        Blacks
Type of sick encounter (N=396)      (n=81)        (n=315)      P value

Mean number of office visits*
  All                                8.49 (6.5)    6.1 (5.0)   < 0.001
  Acute otitis media                 1.75 (2.0)    0.68 (1.0)  < 0.001
  Asthma/RAD                         0.79 (1.2)    0.84 (1.3)    0.720
  RTI                                4.6 (3.6)     3.5 (3.2)     0.007
Emergency room visits
  (% with one or more)
  All ER visits                     23.8          36.2           0.013
  Acute otitis media                 2.0           4.8           0.198
  Asthma/RAD                         5.5           6.7           0.631
  RTI                               16.2          27.3           0.014
Hospitalizations (% one or more)
  All hospital admissions           12.4          10.6           0.366
  T & A (including PET)             15.53          5.46          0.001
Total number of sick encounters*    10.28 (7.26)   7.41 (5.7)    0.01
  (Office + ER + hospitalizations)

*Continuous variables (Standard Deviations).
RAD, reactive airways disease; RTI, respiratory tract infection; T & A,
tonsillectomy and adenoidectomy; PET, pressure equalizing tube; ER,
emergency room.

Table 3. Crude and adjusted relative risks (RR) and 95% confidence
intervals from logistic regression analysis of physician office visits,
ER visits, and hospitalizations by race (whites vs blacks)

                                           RR (95% CI)*
                                           (Controlling for
                         Crude RR          all variables except
Type of Service Used     (95% CI)          private insurance)

Frequent office sick     1.35 (1.22-3.04)  2.45 (1.45-4.03)
Any ER visits            0.89 (0.33-0.96)  0.51 (0.01-0.81)
Any hospital admissions  0.98 (0.90-1.06)  0.8 (0.4-1.6)
Frequent sick visits     1.43 (1.12-1.8)   2.5 (1.3-5.0)
  (Office + ER +
  hospital admission)

                         RR (95% CI)*
                         (Controlling for
                         all variables plus
Type of Service Used     private insurance)

Frequent office sick     2.5 (1.42-3.33)
Any ER visits            0.91 (0.45-1.67)
Any hospital admissions  0.7 (0.3-1.5)
Frequent sick visits     2.17 (1.32-3.58)
  (Office + ER +
  hospital admission)

*Adjusted for covariables of maternal education, maternal age, marital
status, mother's smoking status, mother's employment status, history of
breastfeeding, daycare or preschool attendance, birthweight, age of
child, gender of child, pets at home, and family history of atopy.
ER, emergency room; CI, confidence interval.