OUP user menu

Public and private prenatal care providers in urban Mexico: how does their quality compare?

Sarah L. Barber
DOI: http://dx.doi.org/10.1093/intqhc/mzl012 306-313 First published online: 4 May 2006


Objective. To evaluate variations in prenatal care quality by public and private clinical settings and by household wealth.

Design. The study uses 2003 data detailing retrospective reports of 12 prenatal care procedures received that correspond to clinical guidelines. The 12 procedures are summed up, and prenatal care quality is described as the average procedures received by clinical setting, provider qualifications, and household wealth.

Setting. Low-income communities in 17 states in urban Mexico.

Participants. A total of 1253 women of reproductive age who received prenatal care within 1 year of the survey.

Main outcome measure. The mean of the 12 prenatal care procedures received, reported as unadjusted and adjusted for individual, household, and community characteristics.

Results. Women received significantly more procedures in public clinical settings [80.7, 95% confidence interval (CI) = 79.3–82.1; P ≤ 0.05] compared with private (60.2, 95% CI = 57.8–62.7; P ≤ 0.05). Within private clinical settings, an increase in household wealth is associated with an increase in procedures received. Care from medical doctors is associated with significantly more procedures (78.8, 95% CI = 77.5–80.1; P ≤ 0.05) compared with non-medical doctors (50.3, 95% CI = 46.7–53.9; P ≤ 0.05). These differences are independent of individual, household, and community characteristics that affect health-seeking behavior.

Conclusions. Significant differences in prenatal care quality exist across clinical settings, provider qualifications, and household wealth in urban Mexico. Strategies to improve quality include quality reporting, training, accreditation, regulation, and franchising.

  • Mexico
  • prenatal care
  • quality
  • urban health services

Private for-profit practitioners and pharmacies form a large part of the health market in low- and middle-income countries. Increasing the role of the private sector to meet public policy goals has the potential to mobilize additional health resources, expand access, and promote service responsiveness [1]. Evidence suggests that patients prefer and are satisfied with specific private services [2,3], and poor and wealthy alike use private health care [4]. However, little empirical evidence has demonstrated that strategies to increase the role of the private sector have achieved public policy goals [1,5]. In settings characterized by weak institutional and regulatory frameworks to monitor and evaluate clinical standards, studies have demonstrated that some private for-profit health care providers offer low-quality curative [6–12] and prenatal care [13].

This study examines variations in prenatal care quality by public and private clinical settings and household wealth in urban Mexico. Quality is prominent on the Mexican health policy agenda because of a growing unregulated private sector as well as incentives in the public sector at odds with quality goals. A 2000 national survey reported that 76% of Mexicans thought the health system needed fundamental changes [3]. In response to these concerns, the Ministry of Health launched the Crusade for Healthcare Quality, which seeks to promote technical standards through such means as medical school accreditation and certification of health professionals and facilities [14]. The effectiveness of this campaign on private health care is of particular concern, given that it accounts for more than half of the value of total supply [15].

Despite the policy emphasis, few studies evaluate health care quality in Mexico. In an opportunistic survey of medical doctors treating HIV-positive patients, Bautista-Arrendondo et al. reported 37% adherence to protocols for antiretroviral therapy [16]. Bojalil et al. reported poor management of children presenting with diarrhea and acute respiratory infections among private compared with public practitioners [11]. An ongoing study documents low-quality prenatal care for poor and indigenous women in rural Mexico (SL Barber et al., unpublished observations). Cross-national research has shown substantial variation in the content of routine prenatal care [17,18]. Variations in the content of prenatal care received in less developed settings have been associated with the odds of skilled delivery [19], preterm births [20], perinatal mortality [21,22], and birthweight in grams [23]. This suggests that improving adherence to prenatal protocols in Mexico has important health implications.

This study aims to evaluate whether prenatal care quality varies by public and private clinical settings and by household wealth. Data about urban low-income communities in 17 Mexican states collected in 2003 provide retrospective reports of prenatal care procedures received from 1253 women of reproductive age. From these reports, a 12-item index of prenatal procedures received is generated, and the mean procedures received by clinical setting and household wealth are reported. To ensure that the results reflect differences in care received rather than individual differences, the results are compared with the mean prenatal procedures received adjusted for individual, household, and community factors. The study concludes with recommendations to improve quality in this setting.

The setting

Mexico is a middle-income country of 106 million people, and more than three-quarters reside in urban areas. Health care expenditures amounted to 6% of GDP in 2001 or an average per capita spending of US$ 538 (purchasing power parity); 52% is out-of-pocket [14]. Since the mid-1980s, the Mexican constitution has guaranteed universal health care access.

Health care in Mexico is delivered through social security institutes, Ministry of Health and related government facilities, and private facilities [24]. Social security facilities cover approximately 51% of the population [14], and the largest institutions are the Mexican Institute for Social Security, the Social Security Institute for Civil Servants, the state-owned oil company, the air force, and navy. The mandate of the Ministry of Health and other government facilities is care provision for some 40% of the population without insurance to ensure the constitutional right to universal access.

Despite the extensive public networks, 21% of social security beneficiaries and 28% of non-beneficiaries also use private health care providers [14]. The private sector is composed of a diverse range of providers and services. At the top end of the spectrum are specialist clinics and hospitals that provide internationally competitive standards of care. At the other end are traditional healers and midwives, other non-allopathic providers, prescribing by untrained pharmacy staff, and a small private non-profit sector.


The data are from the 2003 Encuesta de Evaluación Urbana, a survey commissioned by the Mexican government to evaluate the urban Oportunidades poverty reduction program initiated in 2002. The survey used a cluster sampling design based on administrative blocks in low-income urban populations across 17 states. Details of this survey have been documented elsewhere [25]. A total of 16 125 households were surveyed or 90% of targeted households; 2% refused to be interviewed.

The survey’s fertility module collected detailed information about prenatal care procedures received and the source of care from women of reproductive age. Among 1409 women who were pregnant within 12 months of the survey, omitted from the analyses are women who did not seek prenatal care and those who were pregnant at the time of the survey or reported a miscarriage or abortion. The latter groups may not have completed a full series of consultations. With these exclusions, the analyses focus on 1253 women.

The quality of prenatal care is measured by a series of questions about prenatal services received that correspond with national clinical guidelines [26]. The 12 activities are those routinely conducted during history-taking and diagnostics (blood and urine samples, and history of bleeding and discharge), the physical examination (blood pressure and weight, and measurement of uterine height), and other preventive procedures (tetanus toxoid immunization and iron supplements, advice about family planning and breastfeeding, and use of the health card). The 12-item index has a Cronbach alpha scale reliability coefficient of 0.83, indicating a satisfactory level of internal consistency. Similar to previous analyses [23], the 12 procedures are summed up, and the main analyses report the mean prenatal procedures received.

There is good reason to believe that the 12 activities reflect quality and influence health and behavioral outcomes in Mexico. A blood sample and iron supplements are related to detection and prevention of anemia. The prevalence of anemia among pregnant women in Mexico averages 28% [27], and iron deficiency anemia at full gestational age has been reported at 40% or higher [28]. Taking a urine sample and measuring blood pressure may detect hypertensive disorders of pregnancy, which is an important direct cause of maternal mortality [29]. Bleeding and vaginal discharge are danger signs during pregnancy. The importance of a thorough physical examination was demonstrated in research showing associations between a more complete physical and birth outcomes [23]. Practices such as non-sterile cord care have been reported as factors in neonatal tetanus deaths in Mexico [30]. Advice about breastfeeding and postpartum acceptance of family planning promotes maternal and child health and birth spacing. Recording information on a health card promotes care continuity.

The unadjusted proportion for each of the 12 procedures is reported, as well as the average across all procedures. However, maternal reports of prenatal care procedures received provide multiple data about care received in each clinical setting. To account for the possibility that such reports may reflect systematic differences in individual and socioeconomic characteristics, the adjusted mean is also examined. This involves a two-step procedure whereby community fixed effects linear regression models are used to explain the 12-item index on individual and household characteristics. The mean value of the index is then predicted while holding the remaining covariates at their mean values. This result is an estimate of the average quality received purged of differences in individual, household, and community characteristics.

The adjusted means control for a series of individual and socioeconomic characteristics related to maternal risk or socioeconomic factors affecting health-seeking behaviors that vary across the sample. At the individual level, maternal reports of previous birth complications are included because they signal a high-risk pregnancy and may have resulted in a different level of care. Maternal age is expressed as a continuous variable in years and squared to account for non-linearity, given that young and older maternal age is considered high risk. At the household level, the age of the household head in years, and educational levels for the mother and head of household are included. Education is expressed as dummy variables for none, some primary, completed primary, and secondary or more. Other socioeconomic characteristics are solid household wall construction, monthly household consumption, and household size. Solid wall construction is defined as stone, brick, or partition. Household wealth is estimated with a consumption index, which measures monthly household expenditures on food and non-food items. Household expenditures are considered a more reliable measure of income in settings where many people have seasonal or informal employment [31]. Dummy variables identifying the household head’s civil status categorized as civil union, married, or other (single, divorced, separated, or widowed) are included. Lastly, child’s year of birth is included to control for recall bias.

The analyses are presented by the clinical setting where prenatal care was received. The survey identifies 11 settings, which are divided into four major categories: (i) social security facilities; (ii) government services for the uninsured, including Ministry of Health facilities; (iii) private facilities; and (iv) private solo practitioners. The first clinical setting identifies women who are insured. In the analyses, we combine clinical settings into public (social security and government facilities) and private (private facilities and solo practitioners). The survey categories do not distinguish between clinics and hospitals within a given clinical setting. However, the sample is composed of a low-income population, and it is expected that the majority access prenatal care at primary rather than secondary facilities. We also identify two provider qualifications: medical doctors and non-medical doctors. Non-medical doctors in public settings include trained nurses and midwives, whereas those in private settings are trained practitioners in addition to community health workers.

The analyses are divided into four parts. The sample characteristics are first described by the four clinical settings. Secondly, the unadjusted proportion of the 12 prenatal procedures and the mean of the 12 procedures are presented by the four clinical settings. Thirdly, differences in mean prenatal procedures received by provider qualifications are evaluated within clinical settings. The clinical settings are collapsed into public and private to ensure sufficient observations for analysis with each setting. Both unadjusted and adjusted means are reported. For the adjusted means, the linear regression models interact the aggregate public or private clinical setting with provider qualifications, and regressions control for the individual and household factors previously described. Lastly, the unadjusted and adjusted mean procedures received by household wealth are reported within public and private clinical settings. Linear regressions predicting the adjusted means interact public and private clinical settings with household wealth as measured by four quartiles of monthly household expenditures. Given large differences by provider qualification, the adjusted models examining differences by household wealth also control for receipt of care by a medical doctor. Linear trend tests are performed to test wealth discrepancies within clinical settings, by using community fixed effects linear regression models interacting clinical setting with the variable measuring wealth quartile, while controlling for individual and household factors previously described. All analyses were conducted in STATA, and standard errors were corrected for the cluster survey design.


Table 1 describes the sample. The vast majority of women (87%) obtained prenatal care from medical doctors rather than nurses, midwives, or other health care providers. Only 13% of women who sought care from private solo practitioners received care from a medical doctor, however. Significant differences exist in maternal and household characteristics across clinical settings. Women who received care at social security facilities are associated with higher proportions of previous birth complications and married household heads, in addition to higher household head educational levels and expenditures. Women who received care from private solo practitioners are associated with lower rates of previous birth complications, lower household head and maternal educational levels, higher rates of civil union among household heads, and lower expenditures. Some 40% of women who attended private clinics reported previous birth complications. More than half of the sample (59%) sought prenatal care at government facilities, and 73% went to any public facility.

View this table:
Table 1

Characteristics of the sample by clinical setting, urban Mexico

Characteristics of sampleSocial security clinicsGovernment facilitiesPrivate clinicsPrivate solo practitionersTotal observations
Provider qualifications (%)
    Medical doctor97.292.496.212.687.3*
    Nurse, midwife, other providers2.87.63.987.412.8
Maternal and household characteristics
    Previous birth complications (%)34.629.039.722.331.3*
    Maternal age (year)26.725.926.325.326.0
    Age of household head (year)36.536.337.133.336.3
    Education of household head (%)
        Completed primary23.524.325.627.224.7
        Secondary or more19.
    Education of spouse (%)
        Completed primary20.720.622.710.720.2
        Secondary or more8.
    Household head civil status (%)
        Civil union19.634.224.861.232.6*
    Solid household wall construction63.159.473.553.462.1*
        Monthly household expenditures (pesos)3680.02202.33061.01977.62555.3
        Household size5.
Percent by clinical setting (%)14.358.818.78.2100.0
  • * Significant differences at P ≤ 0.05.

Table 2 summarizes the unadjusted proportion of prenatal care procedures received by clinical setting. The lowest scores were for urine sample (62%), blood sample (63%), and recording the information on a health card (63%). The highest scores were being weighed (92%) and having blood pressure taken (91%). Significant and large differences exist, however, across clinical settings for each procedure (P ≤ 0.05). Women who attended social security clinics received the most prenatal procedures (88%), followed by government facilities (80%), private clinics (71%), and private solo practitioners (33%). The two private settings are associated with fewer procedures for all activities, including history-taking and physical examination, as well as those associated with services fees. In separate analyses (not shown), we predict the adjusted means for the individual procedures and the mean for the 12 procedures by clinical setting. In comparing the unadjusted and adjusted means, we find no significant differences at the 5% level.

View this table:
Table 2

Unadjusted percentages and 95% confidence intervals in prenatal care procedures received by clinical setting1

Prenatal care proceduresSocial security clinicsGovernment facilitiesPrivate clinicsPrivate solo practitionersAverage
History and physical examination
    1. Asked about bleeding82.1 [75.7,87.1]71.0 [67.2,74.5]64.5 [56.3,72.0]27.2 [20.8,34.7]67.8 [64.0,71.3]
    2. Asked about vaginal discharge83.2 [78.2,87.3]77.5 [73.0,81.4]71.8 [64.3,78.2]31.1 [23.9,39.3]73.4 [69.5,77.0]
    3. Blood pressure taken98.9 [95.6,99.7]95.5 [93.9,96.8]93.2 [89.3,95.7]40.8 [27.3,55.9]91.1 [86.5,94.2]
    4. Weighed98.3 [95.6,99.4]97.8 [96.5,98.7]90.2 [84.8,93.8]41.8 [29.1,55.6]91.9 [87.8,94.7]
    5. Uterine height measured86.6 [78.7,91.9]83.0 [79.9,85.8]76.5 [69.6,82.3]32.0 [21.1,45.5]78.1 [74.2,81.6]
    6. Blood sample taken86.6 [81.4,90.5]66.6 [62.5,70.5]53.4 [47.0,59.8]17.5 [10.9,26.8]63.0 [58.7,67.0]
    7. Urine sample taken83.8 [77.9,88.4]64.5 [60.8,68.0]58.1 [51.7,64.3]18.5 [10.9,29.5]62.3 [58.3,66.1]
    8. Tetanus toxoid immunization90.5 [85.1,94.1]89.0 [85.8,91.6]72.7 [65.5,78.8]47.6 [36.2,59.2]82.8 [79.4,85.7]
    9. Iron supplements91.1 [84.9,94.9]86.4 [83.3,89.1]80.8 [75.6,85.0]38.8 [25.8,53.8]82.1 [77.4,86.0]
    10. Advised about lactation89.9 [85.7,93.0]88.7 [85.7,91.2]82.1 [76.4,86.6]50.5 [39.2,61.7]84.5 [80.7,87.7]
    11. Advised about family planning84.9 [78.7,89.5]69.9 [65.0,74.3]58.6 [52.0,64.8]33.0 [23.3,44.4]66.9 [62.8,70.7]
    12. Recorded information on health card82.7 [75.7,88.0]70.0 [66.1,73.7]47.0 [40.8,53.4]19.4 [11.0,32.0]63.4 [58.6,67.9]
Average across 12 items88.2 [85.9,90.6]80.0 [78.2,81.8]70.7 [67.6,73.1]33.2 [24.6,41.8]75.6 [72.5,78.7]
  • 1 Chi-squared tests indicate significant differences in frequencies by clinical setting for all procedures (P ≤ 0.05).

Table 3 summarizes the mean of the 12 procedures received by aggregate public and private clinical settings and provider qualifications. Panel A describes the unadjusted means and panel B reports the means adjusted for the individual, household, and community characteristics. Significant differences exist between public and private clinical settings and between medical doctors and nurses, midwives, and other health care providers for all comparisons (P ≤ 0.05). On the basis of the adjusted models in panel B, women who received care in public settings received 80% of the procedures compared with 60% in private settings. Women who received care from medical doctors received 79% of the procedures compared with 50% for nurses, midwives, or other health care providers (P ≤ 0.05). Obtaining care from nurses, midwives, or other non-medical doctors working privately is associated with receipt of only 36% of the procedures. No significant difference exists between the unadjusted and adjusted means.

View this table:
Table 3

Unadjusted and adjusted mean prenatal care procedures received and 95% confidence intervals by public and private clinical settings and provider qualifications

Provider qualificationsPublicPrivateAverage
Panel A. Unadjusted means
    Medical doctor82.3 [80.8, 83.9]71.4 [68.2,74.5]79.9 [78.2,81.7]
    Nurse, midwife, other providers71.3 [65.6,77.0]30.1 [22.0,38.3]45.8 [36.7,55.0]
    Average81.6 [80.0,83.2]59.2 [51.9,66.6]75.6 [72.5,78.7]
Panel B. Adjusted means1
    Medical doctor81.2 [79.8,82.6]70.7 [67.9,73.4]78.8 [77.5,80.1]
    Nurse, midwife, other providers73.2 [68.0,78.3]34.8 [30.5,39.1]50.3 [46.7,53.9]
    Average80.7 [79.3,82.1]60.2 [57.8,62.7]75.6 [74.5,76.7]
  • 1 Panel B means adjusted for maternal and household head educational levels, maternal age and age squared, age of the household head, previous birth complications, household monthly expenditures, female headed households, household size, solid walls, household head civil status as married or in civil union, and child year of birth. No significant differences exist between unadjusted and adjusted means.

Lastly, we examine the unadjusted and adjusted mean procedures received by household wealth quartile (panels A and B, Table 4). It is important to recall that the wealth quartiles are relative to the sample, which is drawn from the poorest segment of the population. On the basis of the adjusted models (panel B), the poorest sample quartile received significantly fewer procedures compared with the wealthiest on average and in private clinical settings (P ≤ 0.05). Linear trend tests performed to test wealth discrepancies (not shown) indicate that an increase by one wealth quartile (measuring from poorest to wealthiest in the sample) is associated with a small but significant increase in procedures received on average [1.3%, 95% confidence interval (CI) = 0.0–2.6; P = 0.05]. However, within private clinical settings, an increase by one wealth quartile is associated with a 5% increase in procedures received (4.7%, 95% CI = 2.42–7.1; P < 0.01). Unadjusted means between the poorest and wealthiest quartiles in public settings are also significant, but these differences disappear after controlling for provider qualifications; and individual, household, and community characteristics.

View this table:
Table 4

Unadjusted and adjusted mean prenatal care procedures received and 95% confidence intervals by public and private clinical settings and household wealth quartiles1

Household wealth quartilePublic2PrivateAverage
Panel A. Unadjusted means
    1 (poorest)79.2 [75.6,82.8]38.7 [24.5,53.0]68.2 [60.6,75.9]
    280.6 [78.1,83.1]58.7 [51.2,66.1]75.4 [72.1,78.7]
    382.3 [79.6,85.0]67.4 [60.7,74.2]78.7 [75.8,81.5]
    4 (wealthiest)84.6 [82.0,87.4]71.0 [65.5,76.5]80.2 [77.4,83.1]
Panel B. Adjusted means3
    1 (poorest)78.6 [75.6, 81.6]53.1 [48.2,58.0]71.6 [68.9,74.3]
    279.0 [76.3,81.7]62.9 [58.1,67.6]74.6 [72.2,76.9]
    379.1 [76.4,81.9]69.2 [64.4,73.9]76.3 [73.9,78.7]
    4 (wealthiest)80.4 [77.4,83.5]69.5 [65.3,73.7]77.7 [75.1,80.2]
  • 1 Linear trend tests indicate significantly higher mean procedures received with increases in wealth quartile on average (P = 0.05) and in private clinical settings (P ≤ 0.05), while adjusting for provider qualifications, maternal and household head educational levels, maternal age and age squared, age of the household head, previous birth complications, household size, solid household wall construction, household head civil status as married or in civil union, and child’s year of birth.

  • 2 Differences within public clinical settings significant for unadjusted means only (P ≤ 0.05).

  • 3 Adjusted means control for all variables listed in footnote 1.


The data represent low-income urban populations across 17 Mexican states and not the country as a whole. Rural health care is analyzed separately. Omitted are women who did not seek prenatal care—a group that may be vulnerable to poor outcomes. Data about gestational age were not collected in the survey; thus, we are unable to evaluate women who experienced preterm birth, which presents a reduced opportunity for care. We assess only the quality of prenatal services, which satisfies the criteria for a tracer service to evaluate overall care quality. However, the quality of preventive services may differ from curative or chronic care. We do not evaluate satisfaction or the lack of medicines—a frequent criticism of the public health system in Mexico [3]. Public and private facilities are not measured comprehensively, and the survey does not distinguish between clinics and hospitals within each clinical setting. However, the sample is composed of a low-income population that would be expected to access routine care at primary rather than secondary level facilities.

The content of prenatal care is derived from maternal reports. Women with negative outcomes or lower socioeconomic status may experience prenatal care or recall to events differently compared women with positive outcomes. Patient reports of health care received usually adjust for case mix, because variation in provider practice may result from severity of illness or socioeconomic characteristics that affect the choice of clinical setting or individual recall. A previous study found that prenatal care may not require adjustment for case mix because the vast majority of procedures are routine [32]. In this research, there were no significant differences comparing unadjusted and adjusted means in simple analyses. However, research from a heterogeneous population in rural Mexico over a longer period (1997–2003) consistently found significant differences between unadjusted and adjusted mean prenatal procedures (SL Barber et al., unpublished observations). The need to adjust for individual and socioeconomic characteristics in the analyses of prenatal care should consider variation in socioeconomic and patient characteristics across the sample and the time span between the health care visit and interview that may affect recall.

More than one-quarter of our sample (27%) sought care in the private sector, where they received significantly fewer prenatal procedures. Previous studies across diverse low-income settings report lower quality privately provided curative [6–12,33] and prenatal care [13] (SL Barber et al., unpublished observations). Other countries have implemented various forms of health regulation and legislation to ensure minimum quality standards. In Mexico, large hospitals participate in voluntary accreditation. However, private facilities are considered ‘entrepreneurial initiatives’ not restricted by regulations about services or quality [15]. Realizing the role of the private sector in meeting public policy goals requires explicit policies, and several strategies applicable to this setting are discussed briefly.

This study finds that women across all four wealth groups received significantly more procedures in public compared with private settings. These differences can be seen across activities related to knowledge, such as history-taking and physical examination, as well as diagnostics that may depend on ability to pay. This suggests that women seeking private prenatal care are choosing to pay for services of low technical value. Quality reporting has been employed in the United States to promote transparency in technical quality and inform care-seeking choices. Recent research finds that quality reports influence employee health plan choice [34]. Barriers for low-income populations in Mexico include low educational levels, limited choice, and adequate data to control for case mix. Quality reporting, however, may promote accountability and transparency among health care organizations.

Previous studies have reported associations between prenatal care quality and household wealth [13]. This study finds that the wealthy received higher quality private care, which may be due in part to the ability to pay or financial incentives. The government’s 2004 health reform aims to overcome the financial constraints in accessing health care based on employment and ability to pay, by creating a voluntary health insurance program. Funded from general revenues and enrollee premiums, it replaces user fees for services and medicines with a heavily subsidized sliding prepayment scale and provides a subsidy for the extreme poor. Participating public and private facilities must be accredited, and any health care provider in the national health system can participate. Attracting the uninsured into accredited facilities is promising in ensuring minimum quality standards, assuming that serious consideration is given to technical quality as part of the accreditation process.

The lowest levels of quality, however, are among solo private practitioners operating outside of formal regulatory, quality, or continuing education systems. Franchising networks have been proposed as a means to improve quality in settings where regulations are weak [4]. Routinely conducted for non-clinical services, franchising involves a comprehensive contract, whereby services or technical support are exchanged under the condition that the franchisee achieves an established standard of quality. While promising as a means to recognize and address low technical quality among informal private providers, more evidence is needed to demonstrate this strategy’s success.

We find that medical doctors are associated with significantly more procedures compared with non-medical doctors across public and private clinical settings. Lower quality prenatal care among nurses and midwives in institutional settings contrasts with international evidence that midwife-managed prenatal care is effective for routine consultations [35]. Upgrading the skills of nurses and midwives in Mexico offers an opportunity to increase technical quality. Given that a large proportion of public health workers also maintain private practices [15], investments in increasing overall clinical capacity for basic services would benefit patients accessing different clinical settings.

In summary, private for-profit practitioners are an important source of prenatal care in urban Mexico. The study finds that private providers and non-medical doctors are associated with receipt of fewer prenatal procedures that correspond to the clinical guidelines. The wealthy received more procedures than the poor from private providers. These differences are independent of individual, household, and community characteristics. Low adherence to evidence-based prenatal protocols has health and cost implications. Strategies to improve quality and advance public policy goals in this setting include quality reports, training, accreditation, regulation, and franchising.


Lisa DeMaria and anonymous reviewers of this journal provided useful comments. This article also benefited from discussions during the analysis of a companion article on prenatal care quality in rural Mexico. The author remains responsible for all errors and omissions. This research was funded by the National Institutes of Health Fogarty International Center. The data were collected by the National Institute of Public Health with funding from the Oportunidades program of the Mexican Ministry of Social Development and the Inter-American Development Bank. No conflict of interest is declared.


View Abstract