The impact of substantial financial incentives on C-section rates: Evidence from Iran

Delivery by cesarean section (C-section) is medically necessary in 10%–20% of births [1, 2, 3]. Unnecessary C-sections result in elevated rates of maternal and infant morbidity (e.g., bleeding, infection, respiratory distress, temporary, and permanent disability) and mortality (i.e., death) [4, 5, 6, 7, 8]. Unnecessary C-sections also result in higher financial costs, including substantial out-of-pocket payments by patients [9]. For these reasons, high C-section rates have long been viewed as a serious public health problem [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15].

C-section rates vary widely across the globe. Data from 154 countries covering almost 95% of world live births show “averages ranging from 5% in sub-Saharan Africa to 42.8% in Latin America and the Caribbean” [7]. C-section rates have also been increasing rapidly everywhere, from 6% of all births in 1990, to 21% in 2018, to an estimated 28.5% in 2030 [7]. A 2018 Editorial in the Lancet summarized the situation as follows: “[T]he global rate of caesarean birth has doubled in the past 15 years to 21%, and is increasing annually by 4%. While in southern Africa use of caesarean section is less than 5%, the rate is almost 60% in some parts of Latin America” [10].

Iran has long had one of the highest C-section rates in the world, so reducing those rates (and the associated maternal and infant morbidity and mortality) is an obvious public health priority [16, 17, 18, 19, 20, 21, 22]. In 2005, 40.7% of births in Iran were by C-section, and in some facilities, the C-section rate was as high as 90% [20]. By 2014, fully 53% of births in Iran were by C-section [12, 21]. In terms of financial costs, the World Health Organization (WHO) estimated that Iran spent more than $100,000,000^.on unnecessary C-sections in 2008—when the rate was materially lower than it is today [9].

During the 2013 Iranian Presidential campaign, reform of the Iranian health system was a prominent issue, with one candidate (Dr Hassan Rouhani, who won the election) placing health at the center of his campaign [23]. In 2014, Iran issued a Health Transformation Plan (HTP) which charged the Iranian Ministry of Health and Medical Education (IMHME) with materially reducing the frequency of C-sections [24]. The reform was motivated by concern about the morbidity and mortality associated with high C-section rates, along with the excessive financial costs being incurred by the Iranian health-care system and patients [23, 24].

Acting in accordance with the HTP, the IMHME modified the payment system to create substantial financial incentives discouraging C-sections. Initially, the reforms only affected public hospitals, but they were subsequently extended to private hospitals.

Table 1 summarizes the total payment and patient share (co-payment) for vaginal deliveries at public and nonpublic hospitals, expressed in terms of the percentage of payment for a C-section performed in the same type of hospital. Table 1 shows these figures at 3 different points in time: pre-reform, post-first reform, and post-second reform.

As Table 1 reflects, before the first reform, physicians who performed a vaginal delivery were paid 81% of amount they would receive had they performed a C-section at the same type of hospital, irrespective of whether the delivery was in a public or nonpublic hospital. The patient’s co-payment (i.e., their share of the payment) was set at 10% of the total payment, so it was also 81% of the amount that a patient who received a C-section would pay.

After the first reform, which took effect on May 5, 2014, physicians who delivered a patient vaginally in a public hospital received a payment that was more than twice as much (i.e., 214%) as the payment for a C-section performed in the same type of hospital. The patient’s co-payment for vaginal deliveries in public hospitals was eliminated entirely. However, as Table 1 reflects, these changes only applied to public hospitals; deliveries at nonpublic hospitals retained the payment arrangement that prevailed before the first reform took effect.

Subsequently, a modified version of these incentives was extended to all hospitals in Iran (second reform). After the second reform, which took effect on September 23, 2014, physicians who performed a vaginal delivery were paid 125% of the amount they would receive had they performed a C-section, irrespective of whether they were doing so at a public or nonpublic hospital. However, patient co-payments varied significantly depending on the type of hospital. Patients who delivered in public hospitals had no co-payment. Patients who delivered in nonpublic hospitals were responsible for a standard share (co-payment) of 10% of the total payment. Because physician payment for a vaginal delivery was set at 125% of the payment for a C-section, this meant that patients who had a vaginal delivery in a nonpublic hospital (paradoxically) paid a higher co-payment than patients who had a C-section in the same hospital.

We exploit this sequential round of reforms to assess the impact of substantial financial incentives on C-section rates using a difference-in-differences (DiD) approach. We use administrative data from the IMHME—including detailed information on every birth in Iran before and after the reforms in question—to analyze the effects of these reforms. Our data cover every birth in Iran from September 23, 2013 (32 weeks before the first reform was instituted on May 5, 2014) to April 21, 2015 (34 weeks after the second reform was instituted on September 23, 2014).

The first reform reduced C-section rates by 4.9% overall, with a larger (8.2%) impact for first-time mothers. By comparing the results from the first and second round of reforms, we show that physician-level financial incentives accounted for 55% of the observed reduction in C-section rates, and patient-level financial incentives accounted for 29% of the observed reduction. There was no evidence of adverse health effects associated with this reduction in C-section rates—and some evidence of health benefits, including lower neonatal intensive care unit (NICU) hospitalizations and fewer maternal deaths.

These findings indicate that financial incentives do have an effect on C-section rates, since they went down by roughly 10% (i.e., a 5% decline on a 53% base). However, these findings also indicate that more aggressive strategies will be necessary to reduce C-section rates to the levels typically recommended by public health authorities (i.e., 10%–20% of births). Stated more concretely, a 60% reduction in Iran’s rate of C-sections is required to get the C-section rate down to 20% of all births, and an 81% reduction is required to get the C-section rate down to 10% of all births—but the substantial financial incentives we study only reduced the C-section rate by 10%.

Section 2 reviews the factors that affect C-section rates and past efforts to reduce them. Section 3 describes our data and provides background on Iran’s health care system. Section 4 outlines our estimation strategy and provides the results. Section 5 considers the policy implications of our findings.

In any given pregnancy, the decision to deliver by C-section is affected by multiple clinical and nonclinical factors. Clinical factors include whether a prior delivery was by C-section, malpresentation, and fetal distress [25]. Nonclinical factors include financial incentives, physician and patient preferences, insurance status, malpractice risk, and hospital characteristics [13, 26, 27]. Physicians may have incentives to prefer C-sections in cases where the clinical indications are marginal; as an article in the flagship Journal of Obstetrics and Gynecology noted:

Another obstetrician put the matter more concisely in an interview with National Public Radio: “[y]ou’re going to pay me more, you’re not going to sue me and I’ll be done in a[n] hour” [29].

In Iran, factors that influence the choice of a C-section are multidimensional, and include economics (e.g., financial, insurance, and medical factors), the organizational and social context of delivery (e.g., the absence of an on-call physician, distrust between obstetricians and midwives), and the personal preferences of the patient (i.e., the perception that vaginal delivery is time-consuming, imposes high-stress levels, and is unpredictable) [30, 31]. Social and cultural factors are also important; a Time magazine article from 2006 noted that “[v]aginal childbirth is very out these days in Tehran .…No longer the provenance of last-minute complications or doctors’ liability fears, caesarean delivery is viewed here as the modern woman’s choice” [20].

Efforts to reduce C-section rates have included a wide array of initiatives, including education (aimed at patients) clinical practice guidelines (aimed at physicians), greater use of midwives, and direct economic incentives (i.e., paying the same amount for a vaginal delivery and a C-section) [13, 14, 32]. The use of financial incentives in this setting has been controversial, and public health personnel have cautioned against their use, except under limited and carefully controlled circumstances [13, 14].

Multiple studies have examined the impact of financial incentives on C-section rates in countries other than Iran [33, 34, 35, 36, 37, 38, 39, 40]. The more sophisticated studies use a DiD approach to compare C-section rates before and after a payment shock and find only a modest (<1%) reduction in C-section rates [32, 33, 36, 37]. Other scholars have also evaluated the impact of the HTP on various aspects of the Iranian health system, including C-section rates [23, 41, 42, 43, 44, 45]. Prior studies focusing on C-section rates have all used an event study approach, and have typically focused on a small number of hospitals [42]. Only one prior study uses nationwide data from Iran, and they find a significant decline in C-sections in public hospitals after the first reform, and in both public and private hospitals after the second reform using an event study approach [41].

Rather than an event study, we use nationwide data from Iran to conduct a DiD analysis, treating each round of reform as exogenous, consecutive, and nonidentical policy interventions. Our use of DiD analysis allows us to separately examine the effects of financial incentives on physicians and patients—unlike previous studies, which used an event study approach. We also examine clinical outcomes in greater depth than previous studies.

The Iranian Maternal and Neonatal database (IMaN) includes detailed information on all births in Iran [46]. IMaN includes information on the date of birth, age of the mother, the number of past deliveries, weeks of gestation at birth, baby weight, birth order (for twins), method of delivery, nationality, number of past pregnancies and abortions, Apgar scores at 1 and 5 min, and information on maternal health and delivery outcomes. The IMaN database also includes information on the province, hospital, and hospital type at which the birth occurred. At each hospital, a specific department is responsible for submitting information to the IMaN database. Multiple previous studies have been based on information from the IMaN database [18, 39].

We cleaned the registry data and dropped observations with maternal age <15 (2,824 births); maternal age >60 (497 births); 13+ prior deliveries (163 births); baby weight >6 kg (74 births); or baby gender unknown (1,112 births). In total, we dropped 4,670 births, or 0.23% of all the births in our dataset. In robustness checks, we obtained similar results to those we report when we included all of these births. We provide additional detail on the demographics of our data in Appendix Table A-1.

As noted previously, the HTP created large financial incentives discouraging the use of C-sections in 2014. Payment to physicians in Iran is calculated by multiplying the “Relative Value Unit” (RVU) for each service by a conversion factor. The United States uses a similar approach to physician payment under Medicare Part B. RVUs and conversion factors are set by an administrative process, which takes account of time and expertise required, location of services, and physician and patient characteristics.¹

The first round of HTP reforms took effect on May 5, 2014, and dramatically increased the RVU for a vaginal delivery performed in a public hospital. The same reform eliminated the patient’s co-payment if they delivered vaginally. Just over 4 months later (September 23, 2014), the IMHME updated the RVUs for both public and private hospitals, which effectively extended a modified version of the first set of reforms to all hospitals.

Table 2 summarizes the applicable RVUs for vaginal deliveries and C-sections during each of the relevant periods (pre-reform, after the first reform, and after the second reform). Each cell is expressed as a multiple of the conversion factor, which is different in public and private hospitals. Apart from these 2 rounds of reform, annual updates in conversion factors affected both physician fees and patient co-payments for vaginal deliveries and C-sections proportionally.

The RVUs in Table 2 are for Iranian mothers; non-Iranian mothers who delivered in public hospitals had a co-payment of 10% of the physician fees (i.e., 2.1k_p prior to the second reform and 4k_p after). Table 2 treats all nonpublic hospitals as if they have the same conversion factor, but there is additional variation based on ownership status (e.g., private, charity, social insurance, and military), which we ignore for purposes of Table 2. No hospital changed its ownership and status from public to nonpublic or vice versa during our study period. Table 1 expresses the same information in Table 2 in terms of the ratio of payment for a vaginal delivery, relative to a C-section.

Finally, we note that the first reform also imposed a physician-level ceiling on payment for C-sections performed in public hospitals. Hospitals are authorized (but not required) to cut payments to physicians whose C-section rate exceeds 45% for any additional C-sections. Unfortunately, we do not have detail on whether and how public hospitals actually implemented this part of the HTP, so we exclude it from Table 2. We treat this aspect of the HTP as a residual factor in our regression analysis in Section 4, after we quantify the impact of the economic incentives imposed on physicians and patients.

Table 3 presents summary statistics for all births in Iran over an extended period (September 2013–April 2015) that includes both rounds of reform, broken down by delivery site (public and nonpublic hospitals). We provide additional detail in Appendix Table A-2.

As Table 3 reflects, 53% of all births in Iran were by C-section, but the rate of C-sections was significantly lower in public hospitals (43%) than in nonpublic hospitals (67%; t-stat: 340). In fairness, as we show in Appendix Table A-2, there is significant variation in C-section rates by hospital type within these broad categories, with some types of nonpublic hospitals having lower rates, and some types of public hospitals having higher rates.

If C-sections in Iran are driven primarily by medical/clinical factors, they should be randomly distributed throughout the week (14.2% per day). Figure 1 shows that although vaginal deliveries are evenly distributed throughout the week, C-sections are not. Delivery by C-section is substantially less common during the weekend (i.e., Thursday and Friday), with the largest difference from the expected rate on Friday (7.3% actual vs. 14.2% expected, t-stat: 85.9).

Similarly, if C-sections are being driven primarily by medical/clinical factors, they should be randomly distributed throughout the day, with 4.1% occurring during each hour. Because of data limitations, we only have the time of delivery for births after March 21, 2015, which is after the second reform has taken effect. However, we have no reason to believe any of the reforms we study affect the timing of a C-section during the day, conditional on a decision to deliver by C-section.

Figure 2 presents the results. We round the time of delivery down to the closest hour—so a delivery that occurs at 7:59 AM is counted as a 7 AM delivery.

As Figure 2 reflects, the period between 8 AM and 1 PM accounts for 50% of all C-sections performed in Iran, even though these time slots only account for 20.8% of the hours in a day. Figures 1 and 2 offer suggestive evidence that many C-sections are not driven by clinical factors but are instead being scheduled for the mutual convenience of physicians and patients.

What about the timing of a C-section relative to weeks of gestation? As Appendix Table A-1 reflects, about 42% of all C-sections are performed at 38 weeks of gestation. A C-section at 38 weeks of gestation is a full week earlier than is commonly recommended, particularly given the known risks of early delivery [47, 48]. In combination, Figures 1 and 2 and Appendix Table A-1 indicate that nonclinical factors substantially influence whether and when a C-section is performed in Iran. Before we turn to formal DiD analysis, Figure 3 visualizes the C-section rate in public and nonpublic hospitals during our study period.

Figure 3 shows that C-section rates in public hospitals decreased substantially after the first reform was implemented, while they remained stable in nonpublic hospitals. The C-section rate in public hospitals declined by almost 7% (significant, t-stat: 167) after the first reform, while it increased by 1.2% in nonpublic hospitals (insignificant, t-stat: 1.9). After the second reform, we observe a downward trend in C-section rates in nonpublic hospitals, while they remained stable in public hospitals. The overall rate of C-sections in Iran dropped by 2.5% if we compare the rate prior to the first reform versus after the second reform. Visual inspection of Figure 3 suggests that there are similar pre-reform trends in C-section rate in public and nonpublic hospitals; we confirm this point using an F-test.²

In Section 4, we use DiD regression analysis to evaluate the impact of these reforms on C-section rates and disentangle the relative impact of the financial incentives created by these reforms on physicians and patients. Our regression analysis controls for various factors, including fixed effects for weeks of gestation, hospitals, and dummies for region and time (weeks of year).

To investigate the impact of the reforms, we estimate the following equation:

y_ijt are our dependent variables of interest including whether a C-section was performed, and maternal and infant health outcomes. D_j is a dummy for hospital, D_t is a dummy for week of the year, D_gt is group time dummy including group trends. In total we have 2,826 dummies in our benchmark regressions.

The unit of observation for our analyses is an individual delivery so we can control for characteristics of mother, baby, as well as labor conditions in X_i. RF is a first reform dummy, and α is our coefficient of interest using a DiD identification. The treatment dummy is 1 in public hospitals and 0 otherwise. Table 4 shows the results, restricted to observations (births) after the first reform, but prior to the second reform.

Table 4 shows that the combined effects of the first reform are statistically significant but economically modest. Iran completely eliminated patient co-payments and increased physician payments by 164% for vaginal deliveries in public hospitals—but C-sections declined by only 4.9% (Table 4, column (6)). In fairness, this is a 10% decline on a C-section base rate of 53%—but Iran still has a long way to go to get to the C-section rates typically recommended by public health authorities (10%–20% of births).

To determine whether there is heterogeneity in the effects of the first reform, Table 5 reestimates Column (6) in Table 4 for various subgroups of mothers and infants.

As Table 5 indicates, we find statistically significant reductions in most subpopulations, although the magnitude of the decline varies. The C-section rate declined across all maternal ages, and baby weights, with the largest decline for the youngest mothers (<20) and babies whose weight ranged from 2.5 to 4 kg. The C-section rate also declined for mothers with health issues that complicate pregnancy (i.e., diabetes and hypertension).

We do not observe a reduction in C-section rates in twin pregnancies—presumably because the decision to perform a C-section in such cases is primarily driven by medical considerations. We observe a smaller absolute reduction among non-Iranian mothers (3.3%), but they have a lower base rate of C-sections than the rest of the population (41% vs. 53%), so the relative reduction is comparable (10% for both groups), even though the co-payment is not waived for non-Iranian mothers who give birth in public hospitals. The largest reduction in Table 5 is for first-born children, where the C-section rate dropped by almost 8%. Vaginal birth after C-section is relatively uncommon in Iran, so the smaller (but still statistically significant) drop of 2.5% for second birth was expected.

Physicians in Iran are required to report to the IMHME the justification for performing a C-section. As we show in Table 6, there are material differences in the reasons given by physicians for performing a C-section in first-time mothers versus mothers who have delivered previously. We focus on the 4-month period immediately after the first reform and before the second reform (i.e., May–September 2014).

As Table 6 reflects, during the period immediately after the first reform went into effect, among first-time mothers giving birth in Iran by C-section, the most popular physician justification were legitimate medical causes (i.e., lack of progress, fetal distress, abnormal presentation, placental/umbilical cord problems, maternal hypertension, cephalopelvic disproportion). These medical causes collectively accounted for 60.1% of births by C-section. By definition, there were no prior C-sections in first-time mothers—meaning that medical causes accounted for 63.5% of C-sections among first-time mothers (60.1% for other medical causes + 3.4% for myomectomy), followed by 18.9% for other (not specified) causes and 17.6% for mother’s request. By comparison, among mothers who had given birth previously, “previous C-section or myomectomy” was the most common reason, accounting for 74.1% of births by C-sections. All other medical causes accounted for 16.4% of births by C-section, followed by other (not specified) at 6.9% and mother’s request at 2.6%. These figures make it clear that the high rate of C-sections in Iran is a complex matter, but many physicians and mothers seem to be adhering to Cragin’s maxim: “once a Cesarean, always a Cesarean” [49].

We now turn to the impact of patient-level incentives on whether a C-section is performed. As Table 1 reflects, the first reform simultaneously increased physician compensation and eliminated patient co-payments for vaginal deliveries in public hospitals. What is the relative contribution of these 2 policies on the overall decline in C-section rates? To analyze that issue, we estimate a second equation that takes advantage of the fact that in the second reform, the patient co-payment for a vaginal delivery in nonpublic hospitals rose, while the co-payment for a vaginal delivery in public hospitals was set at $0 by the first reform. This variation in policies across 2 reforms allows us to estimate the following specification, based on the difference in physician fees and patient co-payments for vaginal delivery versus C-section:

The first term is the nominal difference in physician payment between vaginal and cesarean delivery for each observation. The second term is the nominal gap for the out-of-pocket payment by patient. The difference in physician fees and patient co-payments is expressed in millions of riyals; we obtain similar results when we use fee ratios.

Table 7 reports our estimation of Equation 2, which uses all of our data, including observations after the second reform. Because the patient’s co-payment is 10% of the amount that the physician receives, the magnitude of the coefficients for patient co-payments will be (by construction) 10× the magnitude of the coefficients for physician fees. Similarly, because the second reform dramatically increased physician fees for vaginal deliveries, patient co-payments in nonpublic hospitals went up as well—and exceeded the patient co-payments for C-sections. The result is that the coefficients in Table 7 are negative for physician fees, but positive for patient co-payments.

Table 7 shows that C-section rates are affected by both physician fees and patient co-payments, relative to the fees and co-payments for vaginal delivery. We find that a 1 million Rial increase in the fee for a vaginal delivery (i.e., approximately $40 at the time of the first reform) reduces the C-section rate by 1.1%. The first reform increased the fee for a vaginal delivery in public hospitals by roughly 2.4 million Rials, since the physician fee for a vaginal delivery increased from 17k_p to 45k_p with k_p = 88,000 Rials. This translates into a 2.7% decline in C-section rates, or about half of the observed decline in Table 4.

Table 7 similarly shows that a 1 million Rial increase in the co-payment for a vaginal delivery results in a 9.7% increase in C-sections. In the first reform, the IMHME eliminated patient co-payments, which would otherwise have been 140,000 Rials (i.e., approximately $6 at the time of the first reform). This translates into a 1.4% decline in C-section rates. Putting these 2 findings together, we can explain roughly 84% of the total reduction in C-sections in public hospitals ([2.7+1.4]/4.9) after the first reform, with financial incentives for physicians explaining roughly twice as much of the reduction as financial incentives on patients. The remaining balance (16%) is attributable to factors other than the specific economic incentives imposed on physicians and patients—including any educational initiatives and the impact of allowing public hospitals to cut payments to physicians whose C-section rate exceeds 45% for any additional C-sections.

We used a similar set of regressions to examine whether there is heterogeneity (based on patient-level attributes) in the relative impact of these reforms. As shown in Appendix Tables A-4 to A-6, depending on the demographic group, we can explain 69%–86% of the decline in C-section rates—and 94% of the decline in C-sections at 38 weeks of gestation. Physician fee effects account for two-thirds of the reduction (range: 61%–75%), with patient co-payment effects accounting for the remainder.

What about the health effects of these reforms? Do we find evidence of increased maternal or infant morbidity and mortality? In Table 8 we examine these issues for all births, for all high-risk pregnancies, for all high-risk infants, and for various maternal and infant subgroups. With one exception, we find substantial evidence that the payment reforms in question resulted in either neutral or improved health outcomes for both mothers and babies.

Table 8 indicates that NICU admissions at public hospitals declined for all births, with larger declines for high-risk pregnancies, first and second births, mothers with high blood pressure, and mothers who were 20 years of age or older. Maternal deaths declined as well, as did infant deaths for mothers who were between 20 and 30 years. Maternal ICU admissions declined for all high-risk pregnancies, and for mothers with hypertension, and those giving birth to their second child, but increased for diabetic mothers. Table 5 also indicates that C-section rates for diabetic mothers declined more than for all births. There were no adverse health effects among high-risk infants. The combination of these findings indicates that apart from diabetic mothers, the health impact of the HTP was either neutral or positive. Further investigation of health outcomes among diabetic mothers is obviously warranted.

Reducing high C-section rates is a long-standing global public health priority, although the role of financial incentives in that effort is controversial. For example, recommendations from the WHO on reducing unnecessary C-sections prioritized nonfinancial strategies even while noting “financial incentive[s] remain a major determinant of caesarean births in all settings” [13, 14]. The WHO report cautioned that the use of financial incentives to reduce C-section rates “are recommended only in the context of rigorous research” [13, 14].

We report on the impact of substantial financial incentives on the rate of C-sections in Iran. This exogenous policy reform affected both physician fees and patient co-payments. Although the policy reform initially focused on public hospitals, it was subsequently extended to all hospitals in Iran. Prior to the first reform, physician fees for vaginal deliveries in public hospitals were 20% lower than physician fees for C-sections. After the first reform, physician fees for vaginal deliveries in public hospitals were 141% higher than physician fees for C-sections. Using DiD regression analysis, we show that this policy change caused a 4.9% reduction in the number of C-sections in public hospitals in Iran, with a larger impact on first-time mothers (8.2% reduction).

Iran has also used a variety of nonfinancial reforms to reduce the rate of C-sections, including “mother-friendly hospitals (220 new hospitals offer no-pain delivery methods by the end of 2016); development of standard protocols; preparation classes for mothers (450 thousand classes), midwives, and workshops for specialists and midwives (500 obligatory workshops), and construction of 865 LDR [labor and delivery rooms] in public hospitals” [24]. Other countries have experimented with these and other nonclinical interventions [13, 14, 50]. However, our findings indicate that the economic incentives imposed on physicians and patients accounted for the overwhelming majority of the reduction in C-sections (84%) in Iran.

Of course, the decision to deliver by C-section is affected by multiple factors [51, 52]. And even after the reforms we describe, the C-section rate in public hospitals in Iran is far above that of most countries—and the rate in nonpublic hospitals in Iran is higher still [7, 10]. Our findings also indicate the importance of nonclinical factors on C-section rates, such as the concentration of C-sections on weekday mornings, and at 38 weeks of gestation—a full week earlier than is clinically justified.

These dynamics also point to the limitations of using financial incentives as a “silver bullet” to reduce C-section rates to a reasonable level. Reducing Iran’s C-section rate to 30% (which is still 50% higher than the highest figure recommended in the public health literature [20%]) would require a further increase in physician payments for vaginal deliveries of roughly 550% while keeping C-section payments where they are.

In addition, we note that the formula for the patient co-payment in nonpublic hospitals is working at cross-purposes to the overall goal of reducing C-sections. After the second reform, physicians in Iran are paid 25% more for a vaginal delivery than for a C-section. But, patients who receive care in a nonpublic hospital must pay 25% more for a vaginal delivery than for a C-section, while in public hospitals patients who deliver vaginally have a co-payment of $0. If the goal is to reduce C-section rates, forcing patients who receive care in nonpublic hospitals to pay more for vaginal deliveries than for C-sections makes absolutely no sense. A consistent set of incentives (i.e., doctors get paid more and patients pay less [or nothing] when there is a vaginal delivery) would be a much more sensible design. Indeed, it might make sense to experiment with direct patient stipends for vaginal deliveries, and to selectively target large financial incentives discouraging elective C-sections—that is, those done at 38 weeks of gestation in first-time mothers with normal birth weight babies.

Critically, other than a modest increase in ICU admissions for mothers with diabetes, we find no evidence the HTP had an adverse impact on mothers and infants. Instead, we find evidence the HTP improved birth outcomes, by reducing maternal deaths and NICU admissions.

Because we do not have physician-specific identifiers for the entirety of the period covered by our dataset, we are unable to evaluate physician-level responses to these reforms. Some physicians work in both public and nonpublic hospitals and may have adjusted their practice patterns in response to the first reform. As noted above, we are also unable to quantify the impact of the reform authorizing public hospitals to cut payments to physicians whose C-section rate exceeds 45% for any additional C-sections—partly because we do not have physician-specific identifiers, and partly because we do not know which public hospitals implemented this policy.

Our findings (4.9% reduction in C-section rates from the first reform) are consistent with the only other study to use nationwide data from Iran (6% reduction, using an event study approach) [41]. The consistency of these findings provides reassurance that our results are robust to alternative specification strategies. Our DiD approach also allows us to separate out the impact of the economic incentives created by the HTP on physicians and patients—with the trade-off that unlike an event study, we are not able to quantify the impact of the second reform.

Our findings indicate the WHO might profitably reconsider its advice to prioritize nonfinancial strategies in addressing the high rate of C-sections. Money may not buy happiness, but when appropriately deployed, it can create incentives to do the right thing. That said, our results also show the limits of even substantial financial incentives in reducing C-sections. Other strategies will be necessary to reduce C-section rates in Iran and elsewhere to the levels typically recommended by public health authorities.

We obtained the data from Iran’s Ministry of Health (MOH). We are prohibited from further dissemination, but others can obtain the data from Iran’s MOH.

The authors appreciate the helpful comments that they received when this article was presented at the American Law & Economics Association Annual Meeting in 2022, anonymous comments from 2 reviewers in connection with the 2022 Conference on Empirical Legal Studies, and additional anonymous comments from 2 peer reviewers and the editor at Advances in Global Public Health.

The research was funded by our respective institutions.

None.

All authors were involved in the conceptualization, methodology, and formal analysis. ST was primarily responsible for data curation. ST and MHR were primarily responsible for the original draft. DAH was primarily responsible for review, rewriting, editing, and responding to peer review and editorial comments. Both DAH and MHR supervised ST’s work.

How to cite this article: Hyman DA, Taheri S, Rahmati MH. The impact of substantial financial incentives on C-section rates: Evidence from Iran. Adv Glob Health. 2024;3(1). https://doi.org/10.1525/agh.2024.2317379

Editor-in-Chief: Craig R. Cohen, University of California, San Francisco, CA, USA

Senior Editor: Harsha Thirumurthy, University of Pennsylvania, PA, USA

Section: Ending Poverty