Who is your prenatal care provider? An algorithm to identify the predominant prenatal care provider with claims data.

doi:10.21203/rs.3.rs-3092999/v1

Download PDF

Research Article

Who is your prenatal care provider? An algorithm to identify the predominant prenatal care provider with claims data.

https://doi.org/10.21203/rs.3.rs-3092999/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background: Using claims data to identify a predominant prenatal care (PNC) provider is not always straightforward, but it is essential for assessing access, cost, and outcomes. Previous algorithms applied plurality (providing the most visits) and majority (providing majority of visits) to identify the predominant provider in primary care setting, but they lacked visit sequence information. This study proposes an algorithm that includes both PNC frequency and sequence information to identify the predominant provider and estimates the percentage of identified predominant providers. Additionally, differences in travel distances to the predominant and nearest provider are compared.

Method: The dataset used for this study consisted of 108,441 live births and 2,155,076 associated claims from the 2015-2018 South Carolina Medicaid. Analysis focused on patients who were continuously enrolled throughout their pregnancy and had any PNC visit, resulting in 32,609 pregnancies. PNC visits were identified with diagnosis and procedure codes and specialty within the estimated gestational age.

To classify PNC providers, seven subgroups were created based on PNC frequency and sequence information. Our algorithm was developed by considering both the frequency and sequence information. Percentage of identified predominant providers was reported. Chi-square tests were conducted to assess whether the probability of being identified as a predominant provider for a specific subgroup differed from that of the reference group (who provided majority of all PNC). Paired t-tests were used to examine differences in travel distance.

Results: Pregnancies in the sample had an average of 7.86 PNC visits. Fewer than 30% of the sample had an exclusive provider. By applying PNC frequency information, a predominant provider can be identified for 81% of pregnancies. After adding sequential information, a predominant provider can be identified for 92% of pregnancies. Distance was significantly longer for pregnant women traveling to the identified predominant provider (an average of 5 miles) than to the nearest provider.

Conclusions: Inclusion of PNC sequential information in the algorithm has increased the proportion of identifiable predominant providers by 11%. Applying this algorithm reveals a longer distance for pregnant women travelling to their predominant provider than to the nearest provider. (1)

prenatal care

predominant provider

frequency and sequence

Having a usual source of care has a substantial positive impact on patients and outcomes. Such continuity of care (COC), due to the usual source of care(2,3), is associated with reduced avoidable hospitalization and emergency department visits(4,5), cost reduction (5,6), and reduced probability of mortality.(7,8) Thus, being able to identify a patient’s usual source of care or where they predominantly receive care is important for healthcare resource planning for providers and policymakers.

In the 1990s, algorithms that identified the most frequently visited, or plurality, providers (a provider who provides the most visits) and the majority provider (a provider who provides majority of visits) were introduced using claims data.(9,10) These algorithms used the proportion of visit frequency during a period as well as the specialty of a provider. They have been applied to describe delivery patterns (11) and as measures of care coordination (12). Recently, these algorithms designed for claims data have been verified via electronic health records. The results showed that these algorithms performed better in identifying a primary care physician than a specialist in the primary care setting and that the plurality algorithm outperformed the majority algorithm, as the former can still identify the predominant provider while no provider provided the majority of services.(13)

This plurality algorithm has been applied outside of the primary care setting for patients seeking care for dementia (14) and for those seeking care for HIV(15). However, to the authors’ knowledge, no study has applied this algorithm to identify the predominant provider for prenatal care (PNC). Furthermore, Milton proposed the Adequacy of Prenatal Care Utilization (APNCU) Index specifically for PNC,(16) which has since been widely adopted. The APNCU index integrated the adequacy of initiation of PNC and the adequacy of received services. Therefore, to identify the predominant provider for PNC, an algorithm needs to include both the PNC visit frequency and sequence information, such as initiation and adequacy. However, visit sequential information was not included in previously proposed algorithms.

This study constructed a version of the plurality provider algorithm specifically for PNC by including both the visit frequency and sequence information. This algorithm was applied to pregnant women to identify their predominant providers. As PNC frequency varies, different scenarios were adopted to compare the final percentage of pregnant women whose predominant providers were identified. Finally, this study examined the differences between the travel distance to the nearest visited provider and that to the predominant provider.

Data and sample

This study utilized deidentified Medicaid claims data from the South Carolina Revenue and Fiscal Affairs (SCRFA) Office for 2014–2019. Medical claims were acquired for mothers who had a live birth during the period 2015–2019 in the 12 months prior to delivery. Vital records were used to identify live births with a mom-baby linkage by the SCRFA office. Exemption for this study was obtained from the Institutional Review Board at the authors’ institute due to the secondary analysis using deidentifiable administrative data.

A total of 108,441 live births, with 2,155,076 associated claims, were found during the 2015–2018 period. Full date of birth (DOB) was not released due to privacy concerns. However, as allowed by our data use agreement, and using previously published algorithms as guidance, (17,18) month of birth was estimated by using the dates of service and International Classification of Diseases (ICD) associated with the claims and appended to SC RFA provided year of birth. Samples were delimited to those who had continuous enrollment in Medicaid for the duration of their pregnancies, resulting in 36,848 pregnancies. Among those, only 32,609 included at least one prenatal care claim in the data. Analyses were conducted at the pregnancy level.

Observed prenatal care

This study employed a series of variables to identify PNC visits from these claims. These variables included date of delivery (month/year); date of services (month/year); primary/secondary diagnosis (ICD 9/10 clinical modification (CM), Healthcare Common Procedure Coding System (HCPCS) and Current Procedural Terminology (CPT)); and provider specialty.

Delivery date and services date were used to calculate the duration from each claim to delivery, guided by the previously published algorithm(18). Claims within the gestational age were kept. PNC services were identified with procedure code, including CPT and HCPCS code, and primary diagnosis code. Details can be found in Appendix Table 1.

Table 1

Total prenatal care (PNC) visits and PNC visits at the specialty-specified provider level.
	Total PNC visit	Providers
Provider Counts	N/A	1,922
N	32,609	93,712
Mean	7.86	2.73
Standard deviation	4.12	2.71
PNC > 14 (%)	4.62	0.50
14 ≥ PNC > 8 (%)	40.22	5.07
8 ≥ PNC > 1(%)	48.99	48.51
PNC = 1 (%)	6.18	45.92
Note: N/A-not applicable.

Claims were selected by Medicaid provider specialty and type code, including midwife (06), primary care physician (12,14,19,78), Obstetrics and Gynecology specialist (16, 26, 27), organization (50-FQHC, 97-RHC), nurse practitioner (86), others (02, 10, 40, 48, 57, 94, 95, PA, missing). Specialties that were marked as missing and the supervisor’s specialties of PA (physician assistants) were unknown (0.24%). Although they only accounted for 0.05% of visits, specialties with codes of 02, 10, 40, 48, 57, 94 and 95 were kept for three reasons. First, a coded PNC visit to these specialties satisfies the three criteria of COC.(19) Second, those specialties may serve as a substitute for a traditional PNC provider given the limited access to PNC providers for some populations.(20,21) Finally, pregnancy complications may be different (22) for different women and therefore demand services from different specialties, especially when access to PNC providers is limited.

This study integrates PNC initiation, frequency of visits, and sequence of visits to inform the algorithm. Initiation refers to the provider that provided the initial visit during the pregnancy. Visit frequency at the provider level was then used to estimate the PNC fraction, that is, the percentage of PNC visits to a specific provider, given all PNC visits, for a specific pregnancy. The final PNC visit prior to birth plays a particularly important role in ensuring continuity of care during prenatal, delivery, and postpartum care, so the provider visited at the last PNC visit is considered the referrer for delivery. Therefore, the initial and last PNC visits were identified for each pregnancy.

Provider definition

To identify the predominant provider, this study constructed four categories based on the PNC fraction and sequential information: the exclusive provider (PNC(E)), the majority provider (PNC(M)), the uniquely most frequently visited (PNC(U)) providers, and the most visited providers (PNC(MFV)). The PNC(E)s are those who are the only observed provider throughout the pregnancy, with a fraction of 100%. The PNC(M)s are the providers providing more than half of PNC services for a pregnant woman, with a fraction of larger than 50% - the PNC(M) is aligned with the majority algorithm.(9) The PNC(U)s are providers providing the highest fraction within all visited provider but less than 50% for a pregnant woman - the PNC(U) is aligned with the plurality algorithm.(10) If there were multiple providers who provided equal fractions of PNC visits to a pregnant woman, they were classified as PNC (MFV). Providers who were not classified as any of these categories would not be identified as a predominant provider.

The data obtained from RFA included a unique provider identifier that may refer to a professional or a provider, whichever is applicable, enabling PNC frequency to be counted at the provider identifier level. Since different professionals may share the same provider identifier, frequency can be more accurately counted by combining specialty with this provider identifier. This study used the specialty-specified provider identifier (SSPI) to denote unique providers and calculate PNC frequency.

Different scenarios

This algorithm first counted the PNC frequency at both the pregnancy and provider levels for each pregnancy. The number of PNC providers was also counted for each pregnancy. Every PNC provider was then classified into one of five categories: PNC(E), PNC(M), PNC(U), PNC(MFV) and PNC (non-MFV). The predominant provider was identified initially as PNC(E) through PNC(MFV) as shown in Fig. 1. In the initial scenario, the PNC(E) provider was identified as the predominant provider. Both PNC(E) and PNC(M) were identified as the predominant provider in the second scenario. Beyond PNC(E) and PNC(M), PNC(U) was identified as the predominant provider in the third scenario. The PNC(MFV) was further included as the predominant provider in the fourth, fifth and sixth scenarios. (see Appendix Table 2)

Table 2

The percentage of providers for a pregnant woman at the specialty specified provider level.
Percentage	Definition	Providers
PNC provider counts	All PNC providers	2.87
PNC(E)	Fraction = 100%	28.40
PNC(M)	Fraction (50%, 100%)	19.55
PNC(U)	Largest fraction (N = 1)	33.49
PNC(MFV)	Largest fraction (N > 1)	18.55
Dispersal PNC(MFV)	Largest fraction & COC = 0	7.12
PNC(MFVI)	Largest fraction & 1st visit	7.31
PNC(MFVF)	Largest fraction & last visit	3.06
PNC(MFVIF)	Largest fraction, 1st & last visit	1.42
2	Largest fraction (N = 2)	12.18
3	Largest fraction (N = 3)	3.87
4	Largest fraction (N = 4)	1.41
5	Largest fraction (N = 5)	0.63
6	Largest fraction (N = 6)	0.29
7	Largest fraction (N = 7)	0.12
8	Largest fraction (N = 8)	0.03
9	Largest fraction (N = 9)	0.02
10	Largest fraction (N = 10)	0.00
11	Largest fraction (N = 11)	0.00
Note: n = 32,609. NA-not applicable. PNC(E) – The exclusive PNC provider for one pregnancy. PNC(M) – this provider served more than half of all visits for one pregnancy. PNC(U) – the uniquely most frequently visited provider, who is the only one who served the most visits for a patient (the percentage of this category equals that of PNC(MFV) at count level 1). PNC(MFV) – most frequently visited provider, who served the most visits for a patient. PNC(MFVI) - most frequently visited provider, who served the most visits and the first visit for a patient. PNC(MFVF) - most frequently visited provider, who served the most visits and the last visit for a patient. PNC(MFVIF) - most frequently visited provider, who served the most visits, first and the last visits for a patient. Dispersal PNC (MFV): the number of total PNC visits equals the number of total providers. COC: continuity of care index.

As pregnant women may utilize additional providers that differ from the initial provider accessed during their pregnancy(23), different providers may cover different gestation periods for different women. For example, some of the most visited providers may serve the early pregnancy period, while others may serve later terms. In the early visit (fourth) scenario, the PNC(MFV) provides the initial visit, defined as PNC(MFVI). In the late visit (fifth) scenario, the PNC(MFV) provides the final visit, defined as PNC(MFVF). In the hybrid (sixth) scenario, the PNC(MFV) provides the initial or the final visits. Therefore, PNC(MFV)s in different scenarios may provide care for different needs. The PNC(MFVI) in the early scenario may be associated with PNC initiation, while the PNC(MFVF) in the late scenario may be associated with actual delivery. The PNC(MFV) in the hybrid scenario may be associated with either or both needs. If there is a tie between PNC(MFVI) and PNC(MFVF), this algorithm gave priority to PNC(MFVI) because APNCU listed PNC initiation as one major measurement. (16) The algorithm developed here will include all three scenarios for the PNC(MFV). The percentage of pregnant women with a predominant provider was reported for all six scenarios.

A special case occurs when a pregnant woman visits a different provider at each PNC visit. That means that the number of total visits equals the number of providers. The dispersed nature of all visits makes the first and last visits contain no meaningful sequential information. The predominant provider cannot be reasonably identified under such a situation. Predominant providers could not be identified for those pregnancies.

Application example with travel distance

The results of this algorithm were then applied to examine travel distance differences between the nearest PNC provider and the identified predominant PNC provider. For simplicity, this was delimited to visits that occurred within the state boundaries. Distance was estimated using Google Maps, which calculated the distance from the pregnant woman’s zip-code of residence to the provider’s billing zip-code, measured as centroid to centroid. In cases where the zip-codes were the same, the radius of the zip-code area was estimated using the formula area = π r², where the zip-code area was acquired from the 2010 census. A paired t test was used to compare the travel distances between the two provider types. All analyses were performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC) at a significance level of 95%.

Table 1 summarizes the total number of PNC visits for all pregnancies and PNC visits at the provider level. The average total number of PNC visits for all included samples was 7.86. At the provider level, the average number of PNC visits was 2.73. The number of total PNC visits was divided into four categories based on expected PNC frequencies and those with only one PNC visit. Of all pregnancies, 4.62% had more than 14 visits, 40.22% had 9–14 visits, 48.99% had 2–8 visits, and 6.18% had only one visit.

Table 2 summarizes the number of PNC providers of different definitions at various levels. The average number of providers per woman was 2.87, with slightly more than 7% of pregnant women experiencing extremely overdispersed PNC visits. Less than 30% of all pregnant women in this sample exclusively visited the PNC(E). Another 20% of all pregnant women had more than half of their visits with a PNC(M), the majority provider. Additionally, one-third of all pregnant women had most of their visits with PNC(U), the plurality provider. For 7.35% of pregnant women, the PNC(MFVI) provided the first PNC visit; for 8.31% of pregnant women, the PNC(MFVF) provided the last PNC visit; and for 1.43% of pregnant women, the PNC(MFVIF) provided both the first and the last PNC visits.

Table 3 presents the percentage of predominant PNC providers, subset by PNC frequency. For all pregnant women, the predominant provider was identified mostly as the order of PNC(U), PNC(E), PNC(M), PNC(MFVI) and PNC(MFVF). Compared with a PNC(M), pregnant women who received inadequate PNC (2 ~ 8 times) were more likely to be identified with a PNC(MFVI) or no identifiable predominant provider (both p < .001) and less likely to be with a PNC(E) (p < .001) or PNC(MFVF) (p = .014). Compared with a PNC(M), pregnant women who received adequate PNC (9 ~ 14 times) were more likely to be identified with a PNC(MFVF) (p = .001) and less likely to be with a PNC(E), PNC(MFVI), or no identifiable predominant provider (all p < .001). Compared with a PNC(M), pregnant women who received more than adequate PNC (14 + times) were more likely to be identified with a PNC(MFVI) (p < .001) and less likely to be with a PNC(E) (p = .004) or no identifiable predominant provider (p < .001). Table 4 summarizes the final percentage of predominant PNC providers in different scenarios based on the results from Table 3. The final percentage ranged between 28.4% and 91.8%, with decreasing confidence.

Table 3

The percentage of pregnancies with identifiable predominant prenatal care providers by prenatal care frequency.
Prenatal care frequency	All providers	PNC(E)	PNC (M) (ref)	PNC (U)	PNC (MFVI)	PNC (MFVF)	PNC (MFVIF)	Dispersal PNC
All frequency	100.00	28.40	19.55	33.49	7.31	3.05	1.42	8.19
PNC = 1	6.18	6.18	0.00	0.00	0.00	0.00	0.00	0.00
8 ≥ PNC > 1	48.99	12.24 ^***	9.01	14.98	4.08 ^***	1.28 ^*	0.70	7.40 ^***
14 ≥ PNC > 8	40.22	8.78 ^***	9.52	16.52	3.03 ^***	1.66 ^*	0.67	0.70 ^***
PNC > 14	4.62	1.20 ^**	1.02	1.99 ^***	0.20	0.12 ^**	0.05	0.09 ^***
Note: PNC(E) – The exclusive PNC provider for one pregnancy. PNC(M) – this provider served more than half of all visits for one pregnancy. PNC(U) – the uniquely most frequently visited provider, who is the only one who served the most visits for a patient (the percentage of this category equals that of PNC(MFV) at count level 1). PNC(MFVI) - most frequently visited provider, who served the most visits and the first visit for a patient. PNC(MFVF) - most frequently visited provider, who served the most visits and the last visit for a patient. PNC(MFVIF) - most frequently visited provider, who served the most visits, first and the last visits for a patient. Dispersal PNC: the number of total PNC visits equals the number of total providers.
: p < .05, : p < .01, **: p < .001.

Table 4

The final percentage of identifiable predominant OBGYNs in different scenarios
Scenarios (providers)	Percentage
1, PNC(E)	28.40
2, PNC(E) and PNC(M)	47.95
3, PNC(E), PNC(M) and PNC(U)	81.44
4, PNC(E), PNC(M), PNC(U) and PNC(MFVI)	88.75
5, PNC(E), PNC(M), PNC(U) and PNC(MFVF)	88.40
6, PNC(E), PNC(M), PNC(U) and PNC(MFVI) or PNC(MFVF)	91.81
Note: PNC(E) – The exclusive PNC provider for one pregnancy. PNC(M) – this provider served more than half of all visits for one pregnancy. PNC(U) – the uniquely most frequently visited provider, who is the only one who served the most visits for a patient (the percentage of this category equals that of PNC(MFV) at count level 1). PNC(MFV) – most frequently visited provider, who served the most visits for a patient. PNC(MFVI) - most frequently visited provider, who served the most visits and the first visit for a patient. PNC(MFVF) - most frequently visited provider, who served the most visits and the last visit for a patient.

This algorithm was then applied to measure the travel distance from pregnant women to providers using scenario 6. The travel distances were compared within the South Carolina boundary, and the sample size was reduced to 29,763 (91.3%). The distance was significantly shorter (19.3 vs. 24.2 miles) for pregnant women traveling to the nearest visited PNC provider than to the predominant PNC provider (mean difference: − 4.8 miles, 95% Confidence interval: − 4.6 - -5.1 miles, p < .0001).

This study developed a practical algorithm to identify the predominant PNC provider based on guidelines and actual utilization patterns. The final percentages of identifiable predominant PNC providers were reported based on various scenarios. Prior algorithms only identified approximately 81% of pregnancies with a predominant provider. After adding sequential information, approximately 92% of pregnancies can now be identified. Except for 7% of overdispersed cases, only 1% of included pregnancies could not be identified with a predominant provider using this algorithm. Applying this algorithm revealed that pregnant women travel a longer distance to their predominant PNC provider than to the nearest visited PNC provider.

The implications for this algorithm are threefold. First, the algorithm integrates visit frequency and sequence information, rather than relying solely on frequency as in previous algorithms.(9,10) A recent study applied three different criteria when both plurality and majority algorithms fail. Sequence (the last visit), expenditure and duration were used as alternatives to solve a tie issue when two providers were equally eligible to be listed as the predominant provider.(13) In this study, the well-accepted APNCU was utilized(16) as a justification for assigning the highest weight to the first visit, a higher weight to the last, and the lowest weight to the other visits. Therefore, the percentage of identified predominant providers increased from 81% in the third scenario to 85% in the fifth scenario and to 92% in the sixth scenario.

In addition, this algorithm incorporates dispersion information as a supplement. In this study, those with overdispersed visits, defined as no identifiable COC provider, were mostly (99.7%) classified as having inadequate PNC, with an average and median number of PNC visits of 3. A previous study demonstrated that dispersion increases as the number of chronic conditions increases.(11) If that is also the case for this study, then these women represent a high-risk group. However, the proposed algorithm in this study, as well as previous algorithms, failed to identify the predominant provider for them. Future studies may explore alternative rules to identify the predominant provider for this group.

Second, the identified predominant providers can be applied to better understand how patients seek PNC, particularly for high-need conditions. In this study, pregnant women bypassed the nearest provider and traveled on average 5 miles further to their predominant provider. Understanding why patients might choose to receive care that creates a larger travel burden is important, as these burdens negatively impact access.(21) It is also important to understand why pregnant women sought care from such a large variety of providers for their PNC.

Third, such an algorithm provides a useful tool for healthcare providers and policymakers alike to enhance resources in underserved areas. By having a richer understanding of where pregnant women seek care, where those providers are located and why women bypass the nearest prover, it is necessary to ensure resources are allocated to maximize accessibility. Once this is better understood, efforts to increase PNC providers in specific areas or provide training to existing providers would be more effective.

Future studies are needed to validate this algorithm with direct interviews of pregnant women or electronic health records to verify their predominant provider. The validity of different scenarios should also be confirmed. For example, researchers may plan to include the most visited provider with the last visit in the predominant providers to detect the referral relationship between PNC providers and delivery hospitals. They should recognize that the referral relationship in claims data was not confirmed. The consistency of these two referral relationships should be verified. These interviews should also inquire about why patients seek care with providers who are further away, accepting additional travel burdens.

The proposed algorithm can only be applied in PNC settings because the sequential information adopted in the algorithm was guided by the APNCU, which gives a high weight to the first visit.(16) Therefore, this algorithm may not be applicable in other care settings or for pregnant women in different contexts.

By adding PNC sequential information in the algorithm, the proportion of identifiable predominant providers has increased from 81–92%, an increase of 11%. Different scenarios covered in this algorithm provide researchers and policymakers flexibility in identifying the predominant PNC providers. Applying this algorithm reveals a longer distance for pregnant women travelling to their predominant PNC provider than to the nearest PNC provider.

The findings reported are solely that of the researcher’s and have not been verified by SCDHHS.

Ethics approval and Consent to participate: Informed consent was waived by the Institutional Review Board at the University of South Carolina as the study was classified as Exempt under OHRP Exempt Category 4- secondary analysis using de-identified administrative data.

Research was conducted under the purview of the University of South Carolina's Institutional Review board which reviewed and approved all study procedures. Additionally, study procedures were reviewed and approved by the South Carolina Revenue and Fiscal Affairs (SCRFA) Data Oversight Council which oversees the collection and release of health information collected under § 44-6-170, South Carolina Code of Law. All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication: Not applicable.

Availability of data and materials: The data that support the findings of this study are available from South Carolina Revenue and Fiscal Affairs Office, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Software programs are available upon request to the corresponding author through email: [email protected]

Conflicts of interest: The authors report no conflict of interest.

Funding information: No funding information to be disclosed.

Authors’ contribution: Songyuan Deng — Conceptualization, Methodology, Software, Writing- Original draft preparation; Samantha Renaud — Methodology, Writing- Original draft preparation and Editing; Kevin Bennett—Supervision, Writing- Reviewing and Editing.

Acknowledgements: We would like to acknowledge support from the South Carolina Revenue and Fiscal Affairs Office by approving and providing us the free access to South Carolina Medicaid Claims data. That support made this study possible.

Deng S, Renaud S, Bennett K. Who is your prenatal care provider? An algorithm to identify the predominant prenatal care provider with claims data. BMC Proc (in press).
Rodriguez HP, Rogers WH, Marshall RE, Safran DG. The effects of primary care physician visit continuity on patients’ experiences with care. J Gen Intern Med. 2007;22(6).
Blewett LA, Johnson PJ, Lee B, Scal PB. When a usual source of care and usual provider matter: Adult prevention and screening services. J Gen Intern Med. 2008;23(9).
Romaire MA, Haber SG, Wensky SG, McCall N. Primary care and specialty providers: An assessment of continuity of care, utilization, and expenditures. Med Care. 2014;52(12).
Nicolet A, Al-Gobari M, Perraudin C, Wagner J, Peytremann-Bridevaux I, Marti J. Association between continuity of care (COC), healthcare use and costs: what can we learn from claims data? A rapid review. BMC Health Serv Res. 2022;22(1):1–30.
Kao YH, Lin WT, Chen WH, Wu SC, Tseng TS. Continuity of outpatient care and avoidable hospitalization: A systematic review. Vol. 25, American Journal of Managed Care. 2019.
Gray DJP, Sidaway-Lee K, White E, Thorne A, Evans PH. Continuity of care with doctors - A matter of life and death? A systematic review of continuity of care and mortality. BMJ Open. 2018;8(6).
Baker R, Bankart MJ, Freeman GK, Haggerty JL, Nockels KH. Primary medical care continuity and patient mortality: British Journal of General Practice. 2020;70(698).
Rosenblatt RA, Hart LG, Baldwin LM, Chan L, Schneeweiss R. The generalist role of specialty physicians: Is there a hidden system of primary care? JAMA. 1998;279(17).
Weiner JP, Parente ST, Garnick DW, Fowles J, Lawthers AG, Palmer RH. Variation in Office-Based Quality: A Claims-Based Profile of Care Provided to Medicare Patients With Diabetes. JAMA: The Journal of the American Medical Association. 1995;273(19).
Pham HH, Schrag D, O’Malley AS, Wu B, Bach PB. Care Patterns in Medicare and Their Implications for Pay for Performance. New England Journal of Medicine. 2007;356(11).
Segal JB, Dugoff EH. Building blocks for measuring care coordination with claims data. Popul Health Manag. 2014;17(4).
Dugoff EH, Walden E, Ronk K, Palta M, Smith M. Can Claims Data Algorithms Identify the Physician of Record? Med Care. 2018;56(3).
Yang M, Chang CH, Carmichael D, Oh ES, Bynum JPW. Who Is Providing the Predominant Care for Older Adults With Dementia? J Am Med Dir Assoc. 2016;17(9).
Turner BJ, McKee L, Fanning T, Markson LE. Aids specialist versus generalist ambulatory care for advanced hiv infection and impact on hospital use. Med Care. 1994;32(9).
Kotelchuck M. An evaluation of the Kessner Adequacy of Prenatal Care Index and a proposed Adequacy of Prenatal Care Utilization Index. Am J Public Health. 1994;84(9).
Gourevitch RA, Natwick T, Chaisson CE, Weiseth A, Shah NT. Variation in guideline-based prenatal care in a commercially insured population. In: American Journal of Obstetrics and Gynecology. 2022.
Sarayani A, Wang X, Thai TN, Albogami Y, Jeon N, Winterstein AG. Impact of the transition from icd–9–cm to icd–10–cm on the identification of pregnancy episodes in us health insurance claims data. Clin Epidemiol. 2020;12.
Saultz JW, Albedaiwi W. Interpersonal continuity of care and patient satisfaction: a critical review. The Annals of Family Medicine. 2004;2(5):445–51.
Daw JR, Hatfield LA, Swartz K, Sommers BD. Women in the united states experience high rates of coverage “churn” in months before and after childbirth. Health Aff. 2017;36(4).
Maldonado LY, Fryer K, Tucker C, Stuebe AM. Driving Time as a Structural Barrier to Quality Prenatal Care [21M]. Obstetrics & Gynecology. 2018;
Law A, McCoy M, Lynen R, Curkendall SM, Gatwood J, Juneau PL, et al. The prevalence of complications and healthcare costs during pregnancy. J Med Econ. 2015;18(7).
Carson MP, Chen KK, Miller MA. Obstetric medical care in the United States of America. Obstet Med. 2017;

No competing interests reported.

Supplementfiles.docx

Download PDF

Editorial decision: Revision requested
19 Feb, 2024
Reviews received at journal
18 Feb, 2024
Reviews received at journal
08 Feb, 2024
Reviewers agreed at journal
29 Jan, 2024
Reviews received at journal
26 Jan, 2024
Reviewers agreed at journal
22 Jan, 2024
Reviewers agreed at journal
16 Jan, 2024
Reviewers agreed at journal
05 Jan, 2024
Reviewers invited by journal
05 Jan, 2024
Editor assigned by journal
29 Sep, 2023
Editor invited by journal
25 Jul, 2023
Submission checks completed at journal
25 Jul, 2023
First submitted to journal
21 Jun, 2023

You are reading this latest preprint version

Who is your prenatal care provider? An algorithm to identify the predominant prenatal care provider with claims data.

Status:

Version 1

Abstract

Figures

Background

Method

Data and sample

Observed prenatal care

Provider definition

Different scenarios

Application example with travel distance

Results

Discussion

Conclusion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1