• No results found

Primary Care Physician Supply and Children's Health Care Use, Access, and Outcomes: Findings From Canada

N/A
N/A
Protected

Academic year: 2020

Share "Primary Care Physician Supply and Children's Health Care Use, Access, and Outcomes: Findings From Canada"

Copied!
10
0
0

Loading.... (view fulltext now)

Full text

(1)

Primary Care Physician Supply and Children’s Health

Care Use, Access, and Outcomes: Findings From Canada

WHAT’S KNOWN ON THIS SUBJECT: Results of numerous studies have shown the value of primary care in improving health outcomes. Little is known about the relationship of local primary care supply and access, use, and outcomes of health care services for children under universal health insurance.

WHAT THIS STUDY ADDS: Under universal insurance there are still important differences in primary and ED care use and preventable admissions related to local physician supply. Physician distribution is a critical issue to address in policies to improve access to primary care.

abstract

OBJECTIVES:To describe the relationship of primary care physician (PCP) supply for children and measures of health care access, use, and outcomes.

METHODS:We conducted a population-based, cross-sectional study of all Ontario children from 2003 to 2005. We used health administrative data to calculate county-level supply (full-time equivalents [FTEs]) of PCPs. We modeled the relationship of supply to (1) recommended pri-mary care visits, (2) emergency department (ED) use, and (3) ambula-tory care–sensitive condition admissions and adjusted for neighbor-hood income. We used population-based surveys to describe access.

RESULTS:The county-level PCP supply ranged from 1720 to 4720 chil-dren per FTE. Of the chilchil-dren, 45.4% live in the highest-supply areas (⬍2000 children per FTE) and 8% in the lowest-supply areas (⬎3000 children per FTE). Compared with high-supply counties, the lowest had significantly lower rates of primary care visits (2716 vs 7490 per 1000) and higher proportions of newborns without early follow-care (58.2% vs 14.5%). Low-supply areas had higher rates of ED visits (440 vs 179 per 1000) and admissions. A stepwise gradient existed for every de-crease in supply for most measures. Self-reported access barriers were most evident in areas with⬎3500 children per FTE (32.8% without a physician).

CONCLUSIONS:Under universal insurance there are differences in ac-cess to, and outcomes of, primary care related to local physician sup-ply after controlling for neighborhood income. The most pronounced effect is on primary and ED care use, but there are implications for acute and chronic disease control. Physician distribution is a critical issue to address in policies to improve access to care.Pediatrics2010; 125:1119–1126

AUTHORS:Astrid Guttmann, MDCM, MSc,a,b,c,dScott A.

Shipman, MD, MPH,e,fKelvin Lam, MSc,aDavid C.

Goodman, MD, MS,e,fand Therese A. Stukel, PhDa,d

aInstitute for Clinical Evaluative Sciences, Toronto, Ontario,

Canada;bDivision of Pediatric Medicine, Hospital for Sick

Children, and Departments ofcPediatrics anddHealth Policy,

Management, and Evaluation, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada;eDartmouth Institute for

Health Policy and Clinical Practice, Hanover, New Hampshire; andfDepartment of Pediatrics, Dartmouth Medical School,

Hanover, New Hampshire

KEY WORDS

primary care physician supply, ambulatory care sensitive conditions

ABBREVIATIONS

PCP— primary care physician ED— emergency department

ACSC—ambulatory care–sensitive condition PCAS—Primary Care Access Survey GP— general practitioner FTE—full-time equivalent DA— dissemination unit CI— confidence interval ARR—adjusted rate ratio

The views in this article are those of the authors and do not necessarily represent the views of any funding source. No endorsement by the Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-term Care is intended or should be inferred.

www.pediatrics.org/cgi/doi/10.1542/peds.2009-2821

doi:10.1542/peds.2009-2821

Accepted for publication Feb 10, 2010

Address correspondence to Astrid Guttmann, MDCM, MSc, Institute for Clinical Evaluative Sciences, 2075 Bayview Ave, G106, Toronto, Ontario, Canada M4N 3M5. E-mail: astrid.guttmann@ices.on.ca

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright © 2010 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE:The authors have indicated they have no financial relationships relevant to this article to disclose.

(2)

health outcomes of various popula-tions.1–4Despite this evidence, effective

primary care physician (PCP) work-force distribution remains a problem in both the United States and Canada. Although physician supply has been in-creasing in the United States,5–8 the

PCP workforce for children varies by more than sixfold across primary care service areas, and nearly 1 million chil-dren live in areas without physicians.9

Unlike in the United States, all prov-inces in Canada have a system of uni-versal insurance coverage; however, Canada also has a comparably uneven distribution of PCPs.10,11 Current

esti-mates suggest that 5 million Canadi-ans do not have a designated physi-cian.12 These maldistributions may

lead to inequities in care for children. In the United States, maldistributions are above and beyond those related to a lack of universal health care cover-age. In anticipation of efforts to achieve universal coverage of children in the United States, it is important to understand the degree to which varia-tions in the local supply of PCPs affect access and use of care.

Because of the Canadian health care system’s universal insurance cover-age and the attendant population-based system of data on health ser-vices use, the system provides the opportunity to carefully examine the effect of local primary care supply. Of the 10 provinces in Canada, Ontario has the largest population (12 of 36 million citizens) and is the largest pro-ducer of physicians (6 of 17 Canadian medical schools).

In this study we examined variations in local PCP supply in Ontario and report here the effect of per-capita supply on important measures of care delivery. These measures include self-reported access to care, use of primary and pre-ventive services, emergency

depart-conditions (ACSCs). We also examined the rates of use for several common discretionary conditions.

METHODS

Overall Design

We used multiple, linked administra-tive health data sets from 2003 to 2005 as well as survey, regional resource, and neighborhood-level census data to define and model the relationship of area-based PCP supply, health care use, and acute care use for all children in Ontario. The data are available at the Institute for Clinical Evaluative Sci-ences through an agreement with the Ontario Ministry of Health and Long-term Care. A scrambled health care number allowed for record linkage.

The study received ethics approval from the Sunnybrook Health Sciences Centre Research Ethics Board.

Data Sources

We used the Ontario Health Insurance Plan physician billings to describe phy-sician supply and assess outpatient care. Inpatient and emergency records are available from the Canadian Insti-tute for Health Information’s Dis-charge Abstract Database and the Na-tional Ambulatory Care Reporting System, respectively. The Registered Person Database contains demo-graphic data that include the postal codes of all Ontario residents eligible for health benefits. We used 2 chronic disease databases, the Ontario Diabe-tes Database and the Ontario Asthma Surveillance Information System, to define children with these conditions. These databases’ administrative data algorithms validated against chart re-view allow for high sensitivity and specificity for disease ascertain-ment.13,14 We used waves 2 to 6 of a

2006 cross-sectional household survey (the Primary Care Access Survey

level income quintiles and for popula-tion estimates.

Physician Supply

In Ontario, both general practitioners (GPs) and pediatricians provide pri-mary care to children.15Some

pediatri-cians provide only consultative care, and others have a mixed primary and consultative practice. We used physi-cian billing claims to define the pri-mary care activity for children in each of the 49 counties in Ontario. As in pre-vious work,15 we used an

expert-consensus definition of primary care pediatricians (those withⱖ10 billings annually for well-infant or annual ex-amination care) and defined primary care GPs as those with outpatient bill-ings (including offices in hospitals). We multiplied the proportion of overall billing activity for primary care for children aged 0 to 17 years16by the

physician’s overall full-time equiva-lents (FTEs). FTEs were calculated by using a standard Health Canada for-mula derived from modeling and vali-dation studies that designates physi-cians as 1.0 FTE with 60% of the average billings for that specialty.17

These FTEs were then allocated to the patient’s county of residence so that the physician activity was counted in the county of the patient rather than of the physician (if these differed). Fi-nally, the adjusted allocation was dardized by age-gender indirect stan-dardization by including the 2003 intercensal population estimates. We divided the 49 counties into 5 supply categories in increments of 500 chil-dren per primary care FTE (starting at 1500 children per FTE). These 5 catego-ries were then the main exposure vari-able of interest.

Outcomes

(3)

self-reported access to primary care, ED visits, and admissions for an ACSC. Apart from the survey data, all out-comes were measured with data from April 1, 2003, to March 31, 2005.

Primary Care Use

We divided primary care visits into pre-ventive (well-infant visits, annual ex-aminations, and immunization visits) and all others. We specifically as-sessed the newborn visit in the first week after birth-hospitalization dis-charge for those infants born of nor-mal weight (⬎2500 g) after vaginal de-livery with no maternal or infant complications, to match Canadian follow-up recommendations.18To test

whether areas of relatively high pri-mary care supply might have “exces-sive” visits of a relatively discretionary nature, we defined 3 potentially supply-sensitive types of visits, namely, visits and revisits within 1 week for upper respiratory tract infec-tions, and visits for acne.

Primary Care Access

We used responses from all house-holds with children aged 0 to 15 years (N ⫽ 1042) to 2 questions from the PCAS: “Do you have a family primary care physician?” and for those who did, “Did you have difficulty accessing your primary care physician in the last year?” To account for the design effect of the number of children within the household, as well as the ratio of county to sampling population, each respondent was represented by a pro-vincial weight that reflected the prob-ability of that person being selected for the survey. We then used post-stratification weights that accounted for the total provincial weight to county-age-group1–4,5–9,10–15 population ratio.

The resulting population weight/ county-age-group category was used for estimating the distribution of the survey responses.

Acute Care Use

We assessed all ED use, as well by se-verity, by using the validated Canadi-an Triage Assessment Score.19 We

grouped according to triage catego-ries: acute (urgent to resuscitation) and less acute (semiurgent and nonur-gent). In Canada, patients are not re-quired to see their PCPs before seek-ing ED care. We analyzed ED visits and admissions for ACSCs by using the re-cent pediatric-specific definitions from the Agency for Healthcare Research and Quality.20These area-based

mea-sures include asthma, diabetes, and a group of acute conditions: dehydra-tion, gastroenteritis, bacterial pneu-monia, urinary tract infection, and per-forated appendix. We slightly refined the Agency for Healthcare Research and Quality specifications for asthma and diabetes because we could define these prevalent populations. We did not use age restrictions for diabetes because they were intended to avoid counting incident admissions. Instead we counted events only in the preva-lent population. For asthma we as-sessed events in those aged 2 to 17, because the diagnosis is less clear in children younger than 2 years. We also denominated these admissions on the prevalent asthma rather than general child population, because this method has been shown to improve the preci-sion of comparing regional differences in admissions for chronic diseases.21

We analyzed the acute ACSCs (apart from perforated appendix) as a group.

Covariates

We adjusted all of our health care use outcomes by neighborhood income quintile. We used 2001 Canadian cen-sus mean neighborhood income quin-tile at the dissemination unit (DA). DAs are small (average population of 400). The quintiles are adjusted for house-hold and community size. As such, it is impossible to report the mean income

of a quintile because it varies across areas. Canada does not define a “pov-erty level.” However, these income quintiles are operationalized by Statis-tics Canada to represent “low-income cutoffs,” with an emphasis on the rela-tive cost of living and household size.22

Income is not reported for a small num-ber of DAs (quintile designated missing) with unstable populations (eg, DAs with chronic care institutions or university student housing). For all models with ad-missions as the outcomes, we controlled for regional bed supply.

Analyses

We used logistic and Poisson regression to model the relationship of area-level primary care supply and outcomes. For the Poisson models, the offset was as-signed to stratum-specific cohort popu-lations. For computational efficiency, the unit of analysis was age group (0 – 4, 5–9, 10 –14, and 15–17), gender, and DA stra-tum. Rate and odds ratios and the re-spective 95% confidence intervals (CIs) were obtained by exponentiating the pa-rameter estimates. All analyses were conducted by using SAS 9.1.3 (SAS Insti-tute, Inc, Cary, NC).

Sensitivity Analyses

We tested the robustness of our re-sults with a number of sensitivity anal-yses. Because supply definitions were based on primary care visits, we mod-eled the primary care visit outcomes by supply categories without adjusting for migration (ie, unallocated supply). To assess whether higher numbers of admissions in lower supply areas were related to a lower threshold to admit children who reside further from the hospital, we performed a sub-group analysis of urban children only. Finally, we tested whether trends in asthma and diabetes outcomes were related to the ratio of pediatric special-ists (respirologspecial-ists and allergspecial-ists for asthma; endocrinologists for

(4)

tes) by adjusting for the area-level ra-tio of specialist to PCP supply.

RESULTS

Table 1 lists the distribution of chil-dren, PCPs, and counties according to supply category. Overall, there were 9912 GPs and 362 pediatricians who provided primary care to nearly 3 mil-lion children. Counties with the highest supply of physicians had the highest median household income, with much less pronounced differences accord-ing to area after the 2000 children per FTE threshold.

Fig 1 shows the relationship of missed recommended-visit rates and

self-reported difficulties with access to pri-mary care according to area-level sup-ply. For all 5 measures, there is a trend toward more missed visits and less

ac-cess with every categorical decrease in the local physician supply. Com-pared with the highest-supply

coun-ties, those with the lowest physician

supply (⬎3500 children per FTE) had a

significantly higher proportion of

chil-dren with no PCP (32.8% vs 6.3%) and newborns without follow-up (58.2% vs

14.5%). Table 2 lists the adjusted odds of these missed recommended visits according to physician supply area

and neighborhood income, for which all analyses controlled. For all 3 visit types, there are clear and significant gradients in rates according to supply,

with children in the lowest-supply ar-eas having significantly higher odds of

no preventive or primary care visits

over 2 years and no newborn

follow-up. There were gradients for missed

recommended visits according to neighborhood income for all 3 types of

40 50 60 70 80

Rate

% of children with no preventative care visits over 2 y

% of normal newborns with no follow-up visit in the first week after hospital discharge 4

% of households with difficulty accessing PCPa

0 10 20 30

1500-1999, high supply

2000-2499 2500-2999 3000-3500 >3500, low supply Physician supply as No. of children per FTE

% of children with no primary care visits over 2 y % of households without a PCP

FIGURE 1

Rates of missed recommended services and self-reported access to primary care providers of Ontario children according to area-level physician supply.

aFor respondents with a PCP.

1500–1999 (High) 2000–2499 2500–2999 3000–3499 ⱖ3500 (Low)

Total No. of primary care GPs 5383 2759 1084 497 189

Total No. of primary care pediatricians 271 62 24 4 1

Total No. of FTEs devoted to primary care for children adjusted for migration

741 382 132 46 11

% of primary care FTEs for children by pediatricians 26.6 11.2 13.5 3.9 5.2

No. of counties 7 15 15 8 4

No. of children between 0 and 17 y old 1 366 151 865 685 365 346 152 522 44 458

Median household income, Canadian dollars 62 130 48 834 47 224 46 088 40 766

(5)

visits, although the difference between the lowest and highest income was be-tween 10% and 30%.

Table 3 lists rates of ED and hospital care use according to supply areas. Rates of overall ED use per 1000

increased from 179.8 in the highest-supply areas to 440.3 in the lowest-supply areas. This trend was

pre-dominantly because of low-acuity ED visits. For children with asthma, rates of ED use per 1000 ranged from 18.8 in the highest-supply areas to 50.3

in the lowest-supply areas, again ex-plained largely by low-acuity visits. Rates of asthma and diabetes admis-sions for these children, and acute ACS

admissions for all children, were all higher in the low-supply areas. There was no obvious relationship between ruptured appendicitis and PCP supply.

Table 4 lists the adjusted odds for acute care use and supply and

neigh-borhood income. The most

pro-nounced adjusted rate ratios (ARRs)

according to supply were for low-TABLE 2 Adjusted Odds Ratiosa(95% CIs) of No Physician Visits According to Primary Care Supply

and Neighborhood Income Quintile

No Preventative Care Visits

No Primary Care Visits

No Newborn Visit

Supply

1500–1999 (high) 1.00 (ref) 1.00 (ref) 1.00 (ref)

2000–2499 1.56 (1.54–1.58) 1.27 (1.24–1.31) 2.59 (2.46–2.73)

2500–2999 1.91 (1.87–1.94) 1.42 (1.38–1.46) 3.19 (3.00–3.40)

3000–3500 2.79 (2.70–2.88) 1.69 (1.63–1.76) 3.51 (3.25–3.78)

⬎3500 (low) 5.22 (4.50–6.06) 2.47 (2.14–2.86) 7.44 (6.17–8.96)

Neighborhood income quintile

Missing 1.32 (1.24–1.40) 1.39 (1.26–1.53) 1.31 (1.13–1.53)

1 (low) 1.12 (1.10–1.14) 1.28 (1.24–1.32) 1.37 (1.28–1.47)

2 1.13 (1.10–1.15) 1.17 (1.13–1.20) 1.19 (1.11–1.28)

3 1.08 (1.06–1.10) 1.08 (1.04–1.11) 1.09 (1.01–1.17)

4 1.06 (1.04–1.08) 0.99 (0.96–1.03) 1.01 (0.93–1.09)

5 (high) 1.00 (ref) 1.00 (ref) 1.00 (ref)

ref indicates reference category.

aMain exposure was primary care supply, adjusted for income quintile, age (except for newborn-visit outcome), and

gender.

TABLE 3 Annualized Rates per 1000 Children of Health Care Use and Morbidity Measures According to Area-Level Primary Care Supply (Children per FTE),n

Supply All ED

Visits

Low-Acuitya

ED Visits

ED Visits for Asthmab

Low-AcuityaED

Visits for Asthmab

Asthma Hospitalizationsb

Diabetes Hospitalizationsc

Hospitalizations for Acute ACSCs

Appendectomies for Perforated Appendixd

1500–1999 (high) 179.8 87.6 18.8 3.5 6.2 55.3 2.4 161.0

2000–2499 160.7 89.1 17.0 4.5 8.3 90.5 2.5 136.6

2500–2999 394.5 261.5 35.2 14.8 9.0 91.4 3.0 173.0

3000–3500 361.3 288.7 37.2 22.2 9.1 96.4 3.3 132.5

⬎3500 (low) 440.3 397.8 50.3 37.9 10.1 91.8 4.0 178.6

aCanadian Triage and Acuity Score 4 and 5 (semiurgent and nonurgent). bPer 1000 children with asthma (n560 711).

cPer 1000 children with diabetes (n6686). dPer 1000 appendectomies (n4933).

TABLE 4 ARRsa(95% CIs) for ED Use and Hospitalization According to Area-Level Primary Care Supply and Neighborhood Income Quintile

All ED Visits

Low-Acuity ED visits

ED Visits for Asthma (n⫽560 711)

Low-Acuity Asthma ED Visits

Asthma Hospitalizations

(n⫽560 711)

Diabetes Hospitalizations

(n⫽6686)

Acute ACSC Hospitalizations

Perforated Appendix (n⫽4933)

Area supply (children per FTE)

1500–1999 (high) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 2000–2499 0.79 (0.74–0.84) 0.88 (0.82–0.95) 0.88 (0.82–0.95) 1.24 (1.11–1.37) 1.37 (1.27–1.48) 1.59 (1.25–2.03) 1.02 (0.94–1.06) 0.70 (0.48–1.02) 2500–2999 1.89 (1.78–2.01) 2.54 (2.35–2.74) 1.85 (1.72–1.98) 4.19 (3.77–4.66) 1.62 (1.47–1.78) 1.64 (1.22–2.23) 1.28 (1.20–1.38) 0.78 (0.53–1.13) 3000–3500 1.87 (1.70–2.07) 3.36 (3.02–3.73) 1.98 (1.77–2.22) 6.24 (5.42–7.18) 1.57 (1.37–1.79) 1.81 (1.18–2.77) 1.34 (1.22–1.47) 0.90 (0.61–1.33)

⬎3500 (low) 2.07 (1.70–2.53) 4.36 (3.56–5.34) 2.62 (2.14–3.20) 10.39 (8.41–12.84) 1.65 (1.28–2.12) 1.41 (0.7–2.85) 1.65 (1.34–2.04) 0.72 (0.47–1.10) Neighborhood

income quintile

Missing 0.15 (0.07–0.34) 0.03 (0.00–10.60) 0.71 (0.59–0.85) 0.88 (0.69–1.14) 1.11 (0.84–1.45) 1.02 (0.48–2.18) 1.38 (1.11–1.71) 0.28 (0.19–0.41) 1 (lowest) 1.08 (0.99–1.18) 1.14 (1.04–1.26) 1.00 (0.91–1.09) 1.14 (0.99–1.30) 1.30 (1.17–1.44) 2.03 (1.5–2.75) 1.16 (1.04–1.23) 0.88 (0.73–1.04) 2 1.03 (0.95–1.13) 1.11 (1.01–1.21) 0.97 (0.89–1.06) 1.09 (0.95–1.24) 1.20 (1.08–1.34) 1.39 (1.07–1.8) 1.20 (1.12–1.30) 0.81 (0.68–0.96) 3 0.96 (0.88–1.05) 1.01 (0.93–1.11) 0.94 (0.86–1.03) 1.05 (0.91–1.21) 1.18 (1.07–1.31) 1.30 (0.99–1.72) 1.17 (1.08–1.27) 1.06 (0.90–1.25) 4 1.07 (0.97–1.17) 1.10 (1.01–1.21) 1.02 (0.93–1.11) 1.16 (1.00–1.34) 1.07 (0.96–1.19) 1.38 (1.02–1.88) 1.14 (1.05–1.25) 1.21 (1.03–1.42)

5 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

ref indicates reference category.

aMain exposure was primary care supply, adjusted for age, gender, income quintile, and regional bed supply (hospitalization outcomes only).

(6)

acuity ED visits overall (ARR: 4.36 [95% CI: 3.56 –5.34]) and asthma (ARR: 10.39 [95% CI: 8.41–12.84]). Children with asthma and diabetes who lived in any of the areas withⱖ2000 children per FTE were more likely to be hospitalized than those who lived in areas with ⬍2000 children per FTE. There were small but significant gradients by neighborhood income for low-acuity ED visits, and all ACSC hospitalizations. For instance, children with diabetes who lived in the lowest-income neigh-borhoods were twice as likely as those who live in the highest-income areas to be hospitalized (ARR: 2.03 [95% CI: 1.5– 2.75]). However, physician supply was the more pronounced and consistent driver of the outcomes assessed in our analyses. Table 5 shows a consistent downward gradient of discretionary visits with decreasing physician supply.

Sensitivity Analyses

The use of primary care supply unad-justed for migration attenuated slightly the effect on primary care visit rates for all supply categories, but the gradient according to supply per-sisted. The risk ratios for asthma ad-missions were higher when using un-allocated supply. Restricting the

asthma and acute ACSC hospitalization analyses to urban children had no ap-preciable effect on the supply gradi-ent, except to make the lowest-supply category estimates nonsignificant be-cause there were small numbers. The ratio of 1 specialist FTE per 100 PCP FTEs was protective for asthma admis-sions (rate ratio: 0.98 [95% CI: 0.98 – 0.99]) but did not appreciably change the rate ratios or gradients associated with PCP supply. The specialist ratio for diabetes had no significant effect on these admissions.

DISCUSSION

We have demonstrated that in a universal-access health system, higher PCP supply is associated with great-er self-reported access to care, more use of recommended primary care visits, less use of EDs for nonur-gent problems, fewer hospitalizations for common acute infectious condi-tions, and fewer acute exacerbations (as measured by hospitalizations) for children with asthma and diabetes. For most of these outcomes there is no plateau in the effect of increasing sup-ply up to the level of the maximum in Ontario, with 1720 children per FTE. Al-though we found that ED visits were the lowest in the second-highest,

rather than highest-supply areas, this result may be related to different care-seeking behaviors of inner-city popula-tions where primary care supply is highest. For many outcomes, including the 2 that best reflect access to, rather than use of, care (having a PCP at all and difficulties accessing one for those who have one) and for recom-mended primary care visits, the great-est effect of supply is in areas with the lowest supply: those with⬎3500 chil-dren per FTE. This is the level of under-service associated with designation as a health professional shortage area in the United States.23Our study results

lend empiric support to this definition of PCP shortage.

Authors of a few studies have at-tempted to associate PCP supply with hospitalizations for ACSCs, primarily in the adult and elderly populations, with mixed findings.24,25 Laditka et al26

showed a negative relationship be-tween ACSC admissions and PCP sup-ply in nonrural areas. No study to our knowledge has been specific to chil-dren (especially in the description of physician supply), but the Laditka et al study results suggest that the effect of physician supply on ACSC admissions was most pronounced in children,

Infections Respiratory Tract Infections

Rate per 100 Population

Relative Risk (95% CI)

Rate per 100 Population

Relative Risk (95% CI)

Rate per 100 Population

Relative Risk (95% CI)

Physician supply

1500–1999 (high) 60.7 1.00 (ref) 5.6 1.00 (ref) 4.0 1.00 (ref)

2000–2499 43.3 0.72 (0.68–0.75) 3.5 0.86 (0.85–0.88) 3.6 0.87 (0.85–0.90)

2500–2999 32.5 0.52 (0.49–0.55) 2.2 0.71 (0.70–0.73) 2.9 0.67 (0.64–0.69)

3000–3500 24.6 0.38 (0.35–0.40) 1.5 0.62 (0.59–0.65) 2.9 0.64 (0.61–0.68)

⬎3500 (low) 10.6 0.16 (0.13–0.18) 0.4 0.36 (0.32–0.41) 2.2 0.53 (0.48–0.59)

Income quintile

Missing 0.64 (0.42–0.96) 0.95 (0.87–1.04) 0.76 (0.66–0.87)

1 1.03 (0.96–1.12) 1.02 (1.00–1.04) 0.70 (0.67–0.74)

2 1.10 (1.01–1.19) 1.04 (1.02–1.06) 0.80 (0.77–0.84)

3 1.16 (1.07–1.25) 1.04 (1.01–1.06) 0.87 (0.83–0.90)

4 1.27 (1.14–1.42) 1.06 (1.03–1.08) 0.96 (0.92–1.00)

5 1.00 (ref) 1.00 (ref) 1.00 (ref)

ref indicates reference category.

(7)

which may relate to the distribution of ACSCs in children, with more acute in-fectious conditions such as pneumo-nia and gastroenteritis that may be most sensitive to access to timely care.

Our finding that even in a universally insured population there are small but persistent gradients in ACSC outcomes according to neighborhood income is consistent with those of other Cana-dian studies.27,28However, results of 1

of these studies28suggested that

be-cause outpatient visit rates for ACSCs were highest in low-income popula-tions, differences in these admissions may not be explained by access to care. However, our findings that lower PCP supply is associated with more ACSC ED use and hospital admissions after adjusting for neighborhood in-come suggest that access to primary care does play a role in these out-comes for children. In addition, we have shown that lower-income chil-dren were less likely to have preven-tive and primary care visits, which suggests that worse outcomes for low-income children may be mediated, in part, through health care use, even in the context of a universal-access health care system. However, the overall gradients according to income for most of the outcomes studied were relatively minor, which suggests that although uni-versal coverage cannot eliminate all dis-parities, it can narrow them.

The implications of our work, in both Canada and the United States, should drive policies related to distribution and not merely supply of physicians. Indeed, our finding that discretionary visits increase in concert with physi-cian supply, even in a system that is

relatively undersupplied (as compared with other Organization for Economic Co-operation and Development countries)29,

reiterates conclusions from studies of adults that documented supply-driven demand for health care services30,31and

the importance of physician distribution rather than absolute supply. In the United States, historic increases in both general pediatrician5,32and overall

phy-sician supply have not alleviated physi-cian maldistribution.33 In Canada,

al-though there has been relatively more focus on policies such as financial incen-tives for practice in underserviced ar-eas, and training of international medi-cal graduates in exchange for service in these areas,34the popular debate has

re-mained focused on adding more physi-cians to improve access to care.35

Al-though we could not formally address the question of nonphysician services in this analysis, a fulsome policy debate about improving access to primary care services needs to include consideration of other providers and more innovative models of care, especially in rural ar-eas.36A number of Canadian provinces,

including Ontario, have committed to ex-pand the numbers and role of nonphysi-cian providers such as nurse practitio-ners in primary care delivery.

Although to our knowledge this is the largest, population-based study of the implications of PCP supply for chil-dren, our study had limitations. We did not have data to assess nonphysician care, and although Ontario does not currently have a significant nurse practitioner workforce,34,37it does

pro-vide some primary care services, espe-cially in rural areas. However, our un-dercounting of these services only

affects our estimates of supply of ser-vices rather than of outcomes. Simi-larly, we used a definition of FTE based on Canadian governmental standards. It is possible that we have overesti-mated physician supply with the 60% cutoff; however, this is unlikely to change the association of low physi-cian supply and our outcomes. Finally, our PCAS data were based on adult re-spondents in households with children and, thus, were not specific to the ac-cess to care of the child but of the household, although these two are likely similar.

CONCLUSIONS

Efforts to improve access to care, whether through health insurance re-form in the United States or primary care reform in Canada, need to include policies to address the maldistribution of available health care providers, in-cluding physicians, to achieve the goal of equitable access and outcomes for children.

ACKNOWLEDGMENTS

This study was performed at the Insti-tute for Clinical Evaluative Sciences, which is funded by an annual grant from the Ontario Ministry of Health and Long-term Care). The study was funded by Physician Services Incorporated, and Dr Guttmann holds a salary-support award from the Canadian In-stitute for Health Research.

The opinions, results and conclusions reported in this paper are those of the authors and are independent from all funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred.

REFERENCES

1. Shi L, Starfield B. Primary care, income

in-equality, and self-rated health in the United States: a mixed-level analysis.Int J Health Serv.2000;30(3):541–555

2. Shi L, Macinko J, Starfield B, et al. Primary

care, infant mortality, and low birth weight

in the states of the USA.J Epidemiol Com-munity Health.2004;58(5):374 –380

3. Starfield B, Shi L, Macinko J.

Contri-bution of primary care to health sys-tems and health.Milbank Q.2005;83(3): 457–502

4. Macinko J, Starfield B, Shi L. Quantifying the

health benefits of primary care physician supply in the United States. Int J Health Serv.2007;37(1):111–126

5. Shipman SA, Lurie JD, Goodman DC. The

general pediatrician: projecting future

(8)

6. Goodman DC. The pediatrician workforce: current status and future prospects.

Pediatrics.2005;116(1). Available at: www. pediatrics.org/cgi/content/full/116/1/e156

7. Biola H, Green LA, Phillips RL, Guirguis-Blake J, Fryer GE. The US primary care physician workforce: persistently declining interest in primary care medical specialties.Am Fam Physician.2003;68(8):1484

8. Phillips RL Jr, Bazemore AW, Dodoo MS, Shipman SA, Green LA. Family physicians in the child health care workforce: opportuni-ties for collaboration in improving the health of children.Pediatrics.2006;118(3): 1200 –1206

9. Shipman S, Goodman D, Bethell C, Newton K. Pediatric workforce maldistribution: exam-ining the scope of the problem [abstract]. Presented at Pediatric Academic Societies Meeting; April 29 –May 2, 2006; San Fran-cisco, CA.

10. Chan B.Supply of Physicians’ Services in Ontario. Toronto, Ontario, Canada: Institute for Clinical Evaluative Sciences; 1999

11. Watson D, Black C, Reid R.Primary Health Care in BC: Physician Supply, Distribution and Practice Collocation (1996/97 to 2004/ 05). Vancouver, British Columbia, Canada: UBC Centre for Health Services and Policy Research; 2006

12. Buske L. Projections of physician supply in Canada. Canadian Collaborative Centre for Physician Resources, Canadian Medical As-sociation, May 2009:1– 6

13. To T, Dell S, Dick PT, et al. Case verification of children with asthma in Ontario.Pediatr Al-lergy Immunol.2006;17(1):69 –76

14. Guttmann A, Nakhla M, Henderson M, et al. Validation of a health administrative data algorithm for assessing the epidemiology of diabetes in Canadian children.Pediatr Diabetes.2009;11(2):122–128

15. Guttmann A, Lam K, Schultz SE, Jaakki-mainen L.Primary Care for Children. To-ronto, Ontario, Canada: Institute for Clinical Evaluative Sciences; 2006

16. Jaakkimainen L, Upshur R, Klein-Geltink JE, Leong A, Maaten S, Schultz SE, eds. Primary care in Ontario: ICES atlas. Institute for Clinical Evaluative Sciences. Available at: www.ices.on.ca/webpage.cfm?site㛭id⫽ 1&org㛭id⫽67&morg㛭id⫽0&gsec㛭id⫽0&

17. Canadian Institute for Health Information.

Full-time Equivalent Physicians Report, Canada, 1989/90 to 1993/94. Ottawa, On-tario, Canada: Canadian Institute for Health Information; 1998

18. Facilitating discharge home following a normal term birth: a joint statement with the Canadian Paediatric Society, Society of Obstetricians and Gynaecologists of Can-ada, and Society of Obstetricians and Gynaecologists of Canada. 1996. Available at: www.cps.ca/ENGLISH/statements/FN/ fn96-02.htm. Accessed October 5, 2009

19. Stenstrom R, Grafstein EJ, Innes GD, Chris-tenson JM. The predictive validity of the Ca-nadian Triage and Acuity Scale (CTAS).Acad Emerg Med.2009;10(5):512

20. Agency for Healthcare Research and Qual-ity. AHRQ Quality Indicators. Available at: www.qualityindicators.ahrq.gov. Accessed May 28, 2007

21. Shwartz M, Pekoz EA, Ash AS, Posner MA, Restuccia JD, Iezzoni LI. Do variations in dis-ease prevalence limit the usefulness of population-based hospitalization rates for studying variations in hospital admissions?

Med Care.2005;43(1):4 –11

22. Wilkins R.PCCFVersion 3G User’s Guide (Geocodes/PCCF): Automated Geographic Coding Based on the Statistics Canada Postal Code Conversion Files, Including Postal Codes to June 2001. Ottawa, Ontario, Canada: Health Analysis and Modeling Group, Social and Economic Studies Divi-sion, Statistics Canada; 2001. Catalog No. 82F0086-XDB

23. US Department of Health and Human Ser-vices. Shortage designation: HPSAs, MUAs and MUPs. 2009. Available at: http:// bhpr.hrsa.gov/shortage/hpsacrit.htm. Ac-cessed August 11, 2009

24. Laditka JN. Physician supply, physician di-versity, and outcomes of primary health care for older persons in the United States.

Health Place.2004;10(3):231–244

25. Krakauer H, Jacoby I, Millman M, Lukomnik JE. Physician impact on hospital admission and on mortality rates in the Medicare popu-lation.Health Serv Res.1996;31(2):191–211

26. Laditka JN, Laditka SB, Probst JC. More may be better: evidence of a negative relation-ship between physician supply and

hospi-1148 –1166

27. Agha MM, Glazier RH, Guttmann A.

Relation-ship between social inequalities and ambu-latory care-sensitive hospitalizations per-sists for up to 9 years among children born in a major Canadian urban center.Ambul Pediatr.2007;7(3):258 –262

28. Roos LL, Walld R, Uhanova J, Bond R. Physi-cian visits, hospitalizations, and socioeco-nomic status: ambulatory care sensitive conditions in a Canadian setting.Health Serv Res.2005;40(4):1167–1185

29. Simoens S, Hurst J. The supply of physician services in OECD countries. OECD Health working papers No. 21. Available at: www. oecd.org/dataoecd/27/22/35987490.pdf. Ac-cessed October 5, 2009

30. Fisher ES, Wennberg DE, Stukel TA, Gottlieb DJ, Lucas FL, Pinder EL. The implications of re-gional variations in Medicare spending, part 2: health outcomes and satisfaction with care.

Ann Intern Med.2003;138(4):288 –298

31. Fisher ES, Wennberg JE. Health care quality, geographic variations, and the challenge of supply-sensitive care. Perspect Biol Med.

2003;46(1):69 –79

32. Freed GL, Stockman JA. Oversimplifying pri-mary care supply and shortages.JAMA.

2009;301(18):1920 –1922

33. Goodman DC. Twenty-year trends in re-gional variations in the U.S. physician work-force.Health Aff (Millwood).2004;(suppl Web exclusives):VAR90 –VAR97

34. Tepper JD, Schultz SE, Rothwell DM, Chan BTB.Physician Services in Rural and North-ern Ontario: ICES Investigative Report. To-ronto, Ontario, Canada: Institute for Clinical Evaluative Sciences; 2005

35. Gulli C, Lunau K. Adding fuel to the doctor crisis. Available at: www2.macleans.ca/ 2008/01/02/adding-fuel-to-the-doctor-crisis. Accessed October 5, 2009

36. Goodman DC. Do we need more physicians?

Health Aff (Millwood). 2004. Available at http://content.healthaffairs.org/cgi/ content/full/hlthaff.w4.67v1/DC1. Accessed

October 5, 2009

(9)

DOI: 10.1542/peds.2009-2821 originally published online May 24, 2010;

2010;125;1119

Pediatrics

Stukel

Astrid Guttmann, Scott A. Shipman, Kelvin Lam, David C. Goodman and Therese A.

Outcomes: Findings From Canada

Primary Care Physician Supply and Children's Health Care Use, Access, and

Services

Updated Information &

http://pediatrics.aappublications.org/content/125/6/1119

including high resolution figures, can be found at:

References

http://pediatrics.aappublications.org/content/125/6/1119#BIBL

This article cites 19 articles, 3 of which you can access for free at:

Subspecialty Collections

alth_sub

http://www.aappublications.org/cgi/collection/international_child_he

International Child Health

following collection(s):

This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtml

in its entirety can be found online at:

Information about reproducing this article in parts (figures, tables) or

Reprints

http://www.aappublications.org/site/misc/reprints.xhtml

(10)

DOI: 10.1542/peds.2009-2821 originally published online May 24, 2010;

2010;125;1119

Pediatrics

Stukel

Astrid Guttmann, Scott A. Shipman, Kelvin Lam, David C. Goodman and Therese A.

http://pediatrics.aappublications.org/content/125/6/1119

located on the World Wide Web at:

The online version of this article, along with updated information and services, is

by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Figure

FIGURE 1Rates of missed recommended services and self-reported access to primary care providers of Ontario children according to area-level physician supply.
TABLE 4 ARRsa (95% CIs) for ED Use and Hospitalization According to Area-Level Primary Care Supply and Neighborhood Income Quintile
TABLE 5 Crude Annualized Rates and Adjusteda Relative Risk of Discretionary Visits According to Area-Level Supply and Neighborhood Income Quintile

References

Related documents

Key considerations in the management of head and neck DT are indications for active treatment (such as pain and disease progression in those being surveilled),

– Impact on access to health care for individuals in Texas who receive health care services in Mexico. • Primary care access • Dental services

This study aimed to evaluate the success of ASYCUDA system at the individual level And to determine some factors that can formulate the success of ASYCUDA system. Findings of the

Against such historical and theoretical backgrounds, we fi rst explore the trajec- tory of the Euro-American consciousness and movement from the Atlantic coast to the Paci fi c

At the OECD Ministerial Council Meeting (MCM) of June 2009, ministers from all 30 OECD countries as well as Chile, Estonia, Israel and Slovenia endorsed a mandate for the OECD

Thus the spherical 2-DOF motion of the output link becomes the output of the 2-DOF wrist, and in (2-UU)+U type hybrid sub-chain, two independent base joints in two U joints

“…to help our faculty succeed in their online teaching practice through World Campus and. Continuing Education by identifying and supporting their

Agri-food economics research has to study these specific fields in more depth in order to give useful support to SMEs when they evaluate the economic convenience of