1. Översiktlig projektbeskrivning

Engelsk titel

Medical ecology in a public health care system.
A registry study in Sweden.

Sammanfattning av projektet

Att hälso- och sjukvårdens resurser används på ett optimalt sätt är lika självklart som viktigt. Vid bedömning av optimering är data om sökmönster och patientflöden i vården en grundförutsättning. Med tillgång till grunddata kan statistisk modellering användas för beräkningar av vad som händer i sjukvårdssystemet vid förändringar i vårdbehovet, systemförändringar över tid samt beräkningar av vilka förändringar som kan ge förbättrad optimering av det offentliga sjukvårdssystemet.

Med ovanstående som grund planerar vår grupp vid Primärvårdscentrum, Sundsvalls sjukhus, en studie för att beskriva hur befolkningen i Västernorrland sökt primär- och sjukhusvård under ett år (2006), båda lokalt och regionalt. .

För aktuell studie skall data användas från flera datakällor (Alfa, LISA, Biosis, SYSteamCross, ProfDoc, Swedestar och Obstetrix). Erforderliga tillstånden från varje systemförvaltare har fåtts skriftligt.

Data från de olika journalsystemen exporteras till en oberoende mottagare (Folkhäsocentrum), som via ett scriptprogram (”Avid”, ett befintligt landstingslicensierat system) avidentifierar patienternas personnummer. Avidentifierade data från varje system levereras till studiegruppen.
Ett sekretessavtal upprättas mellan den oberoende mottagaren och studiegruppen, vilket säkerställer att integriteten skyddas.

En gång vi har byggt upp den heltäckande databasen, en ny variabel räknas fram. Den heter "patient-month" inspirerad av de första studierna i Vârdekologi av White et al.
Patient-month mäter antal gånger en individ varit i kontakt med en läkare i sjukvården i en snittlig månad per 1000 innevånnare. Den frambringar en ny aspekt i vårdkonsumtion, nämligen att kontakter med vârden mäts genom en algorytm som väger in periodicitet i kontakten med läkare och totala antal innevånare i Västernorrland 2006. Dessa resultat framställs i "rutor" så som pionjärerna hade gjort i originella artiklar för USA. Man kan då dra nâgra slutsatser om Västernorrlands sjukvårdsystem jämfört med USA:s.

Vi använder multivariat analys för att beskriva sambanden mellan variablen "patient-month" och de olika socio-ekonomiska variablerna registrerade i databasen ang: ålder, kön, adress, nationalitet, civilsttatus och sedvanlig vårdgivarinrättning. Resultaterna är presenterade i Odds Ratio med referensen till den vanligaste variabelnvärden för varje variabeln. Vi beskriver sålunda hur alla personer bakom varje variabelgrupp har använt läkarresurser vid de 5 olika inrättningar i länet: primärvård, akutmottagning, sjukhusinläggning, sjukhus specialist besök, hembesök och Region Universitetssjukhusvistelse.

Vi syftar i denna studie på att få en uppfattning i hur de olika socio-ekonomiska variablerna (individens omgivning) påverkar sjukvårdsanvändning i Västernorrland.

Typ av projekt

Forskningsprojekt

MeSH-termer för att beskriva ämnesområdet

Inlagda MeSH-termer

Academic Dissertations

Works consisting of formal presentations made usually to fulfill requirements for an academic degree.

Hospital Administration

Management of the internal organization of the hospital.

Health Occupations

Professions or other business activities directed to the cure and prevention of disease. For occupations of medical personnel who are not physicians but who are working in the fields of medical technology, physical therapy, etc., ALLIED HEALTH OCCUPATIONS is available.

Health Care Facilities, Manpower, and Services

The services provided in the delivery of health care, associated facilities in health care, and attendant manpower required or available.

Health Facilities

Institutions which provide medical or health-related services.

Health Services Administration

The organization and administration of health services dedicated to the delivery of health care.

Organization and Administration

The planning and managing of programs, services, and resources.

Economics, Hospital

Economic aspects related to the management and operation of a hospital.

Health Care Economics and Organizations

The economic aspects of health care, its planning, and delivery. It includes government agencies and organizations in the private sector.

Economics

The science of utilization, distribution, and consumption of services and materials.

Projektets delaktighet i utbildning

Avhandling
D-uppsats / Magisterexamen
C-uppsats / Kandidatexamen
ST-läkarutbildning
Annan utbildning
Ej del i utbildning

3. Processen och projektets redovisning

Pågående aktiviteter

Planering och förberedelse före datainsamling
Datainsamling pågår
Analys av insamlade data pågår
Författande av skriftlig redovisning / publikation pågår
En eller flera publikationer från projektet är publicerade
Slutfört och inget mer görs inom ramen för detta projekt

Projektstart (när planeringen påbörjas och börjar dokumenteras skriftligt)

2006-05-01

Datum för påbörjande av datainsamling

2006-10-01

4. Detaljerad projektbeskrivning

Bakgrundsbeskrivning

The relationships between people and their health care environments have been known as the ecology of health care, which is a population-based and patient-centred vision of health care that encourages an appreciation of the overall health care use of the community. White defined a model of ecology of health care in 1961 and illustrated it as a series of boxes depicting the proportions of persons participating in health care services in particular settings in an average month (1). This framework has influenced thinking about the organization of health care, medical education, and research since then.

Although much has changed in medicine, organization and financing in the US health care system since 1961 a recent update showed relative stability of estimates of health care use in different settings (2,3,17,18).

It is widely known that health and health care vary substantially with socio-demographic characteristics (4,9,10,11). These variations are potentially important in formulating and assessing health policy objectives, such as access to appropriate care, disease prevention and health promotion, and equity.

In an extension of the health care ecology model considerable variation among subsets of the population was found to be attributable mainly to insurance coverage and usual source of care apart from age, sex, race and ethnicity (3, 10). In these studies the quantification of the variation in health ecology were based on national surveys of households implying methodological weaknesses (2,5,6,19) such as targeting level, recall and report bias and process heterogeneity, which includes different insurance coverage.

No previous study has presented data on an individual level from carefully held central registries in a country with total health insurance coverage.

Syfte

To quantify the variation among subsets of the Västernorrland population receiving care in an average month in five specific care-giving settings.
To account for interactions among socio-demographic, economic and geographic variables and participation in health care to provide information for policy purposes.

Metod: Databearbetning

Analytical Strategy:
We present health care ecology as its pioneers in 1961, featuring how many people by 1000 have participated in health care structures in a typical month during 2006 at the different care levels.
Previous studies performed descriptive analysis according to this Ecology Model 2, using “person-month” records. The “person-month” unit reflects essentially participation in health care instances, rather than frequency or duration of contacts. The value of this record was “0” if a person had no contact with any setting in a month, and “1” if the person had 1 or more contacts in a month.
The monthly “person-month” for each users in year 2006 was added and divided it by 12 for obtaining an estimation of “a typical month”. We then multiplied it by 1000 and divided by the number of persons registered for the correspondent category in the County at the end of 2006. We obtained finally the number of people who visited each setting in an average month for 1000 inhabitants in 2006.
We assessed potentially predictive variables for inclusion in multivariate analysis. Separate logistic regression was used to derive the adjusted OR of the respective predictive variables on the receipt of health care in each setting. The most repeated measure was taken as reference, and the others were related to it.
Statistical analysis was performed with SAS package verison 9.1. The researchers were fully independent of the founders regarding the present work.

Resultat

3.- Results (no figures possible)
Of 1000 persons in Västernorrland, 87 visited a physician’s office in primary care, 44 visited a specialist in outpatient clinics, 20 received care at the emergency department, 14 participated in homecare, 12 were hospitalized and less than 1 was hospitalized in an academic medical centre. 2 persons visited an outpatient clinic at the university hospital.

In table 1 shows the proportion of people receiving care at the different settings. We stratified them by socio-demographic characteristics. In primary care, the youngest infants and oldest adults are attended in a larger proportion than any other group. The same pattern is showed in emergency departments and home care disposals. In outpatient clinics and inpatient regimes the elderly are still overrepresented, but a different distribution in age groups is observed. Women received care in a greater proportion than men in all settings except at the emergency department. It was Swedes participating in health care in a larger proportion than other nationalities. Widowed people received care more often or in more different settings than any other group did. People living in rural areas visited the physicians’ office in a larger extent than those living in urban areas, whereas the latter are more frequently attended a outpatient specialized clinics, emergency departments and inpatient regimes. A larger proportion of home care providing visits is observed in rural areas.

Physician’s office with trust fund has received more people in relation to those under public administration, and people who use to visit those private centres is underrepresented at outpatient clinics. No difference was observed in visits to emergency departments between people listed at public physicians’ office and those at the private centres. Conversely, by 1000 inhabitants, more people who is listed at public institutions is taken care at the university hospital or receiving homecare, in relation to people listed at private centres.

The table 2 shows the results of multivariate analyses in odds ratio for participation in health care. The results are adjusted by socio-demographic characteristics. It is more likely to receive care at any setting when being older than 65 (OR 1.92) or younger than 5 years (OR 2.66). Children have the same probability to be referred to the university hospital than the reference group. Compared to men, women are more likely to attend medical care at any setting except at emergency departments, where they appear to be slightly less likely to receive care (OR 0,94).

At the physician’s office, the OR is 1.5 for women. Swedes and other Scandinavians show a similar relationship in terms of OR for visiting each setting. Larger differences are observed when compared to non-Europeans, who are more likely to enter an inpatient clinic or to receive home care but less likely to be treated on referral at university clinics. Widowed people are more likely to receive care at the emergency departments, in hospital clinics and at home. Married people are less likely to attend at any care setting in relation to divorced or separated. Having a rural residence brings a higher probability for visiting a primary care doctor (OR 1.07) and receiving care at home (OR 1.23), but they are less likely to visit outpatient clinics (OR 0.94) or emergency departments (OR 0.80). Having visited a trusted private medical office is associated with an increased likelihood of receiving care again at the same office, but also with a lower probability of meeting a doctor at outpatient departments (OR 0.87) or at university clinics (0.54).

Diskussion

The “ecology model” highlights the need of comprehensive medical information systems that span all care-providing sites. This framework has repeatedly been a reference for policy makers and educators, since it is based on an overall health care use assessment in a community .
Likewise White et al., we quantify participation as estimate “patient-month” during a year, and we study its relationship with each of the socio-demographic variables. We found this estimate a very useful tool since it is less sensible for an eventual aggregated measure repetition and periodicity in contacts with health care.

Earlier studies are based in national surveys in the USA and UK. They show stability in the proportions of people profiting the system at the different settings through decades, even when taking into account children and dental care. These surveys needed to be adjusted for seasonability (Gallup) or sample size (MEPS) before a multivariate analysis is performed 2. The model may appear to be nested, leading to misinterpretation of “The Ecological Boxes”. Recall bias was minimized in MEPS and Gallup studies using logs 13. In the original article, the authors assess a “need” of health care by asking the population for symtoms. This was out of the scope in our study and is still to be done. However, care chains also imply a “demand” which is specifically generated by the system itself by means of internal procedures (automathically expedited patient citations, controls after surgical procedures, epidemiological follow up, etc). All these visits to a doctor modifies theoretically the participation by increasing the visit rate in a year whithout meaning a larger participation or a broader system coverage. The “person-month” variable applied to a total registry is less volatile at this consideration and brings more information than usual participation variables.

The Ecology Model has never been tested on official registration databases before. A local variant of the same idea was performed in another district in Sweden 14 based upon minor survey data. We think that the application of this model into general, state-administered registries is a useful framework for evaluating equity and adequacy of health care in the society, and has a good reproducibility.

Weighting of our quotients was not necessary, since we took into account all contacts with health system by every individual in this area during 2006. The non-participating individuals were counted from census registry 2006 (Folke), so we reached a 100% rate of continuity: All eligible individuals at the beginning were exactly studied throughout.
Using databases, the recall bias is non-existent and the information bias is minimized by the size of the sample (N) and the variety of people involved in it. The latter could, however, be the source of misclassification. Duplicate registrations were identified and erased by specific orders in program scripting. By means of merging several databases, the probability of contradictory events decreases (16). Matching registries at individual level prevents the risk of some socioeconomic variables to be highly correlated to each other (co-linearity), as they would be if studying clustered individuals instead.

The ecological boxes show a remarkable difference with White’s et al, mostly at primary attention level: there are 217 people seeking care in the US 2001, vs 87 in Sweden 2006. Further, there are 21 individuals-a-month (by average) visiting specialized hospital outpatient clinics in the US, compared to the Swedish 44. This comparison should be done with reservations, as populations and behaviors may be different in Sweden and the US, but there is a probable direct connection between these two proportions for each study. The difference in primary care could partially be explained by the fact of several specialist non-GP having offices at a primary care level in the US in contrast to Sweden 2006, where all primary attention is delivered by GP’s. The proportion for GP-offices in the US if not taking into account outhospital private specialists offices would be 113, much closer to the Swedish 87.

It also appears that Swedish population participates more at the ER level compared to Americans. There is probably a conceptual difference here between studies in defining emergency settings and its availability. The proportion of hospitalized patients in local or academic hospitals, are very similar.

The adjusted OR for participation by the different variables shows that health care system is profited by a greater extent by children and widowed elderly disregarding country of birth. Women appear to participate more than men in all settings but the emergency room. We have not found an explanation for non-Europeans receiving home care to a much larger extent than any other group. This could be generated by small differences in visit registering procedures (“within-group-bias”, misclassification). We were not able to study the impact of incomes and educational level for each person or household.

Due to the geo-demographical characteristics in this county of Sweden, it is completely understandable that a rural residence is associated with more participation in primary care and home care, but less in hospitals.

In this region of Sweden 2006, free election of care provider was not allowed if not changing address to other district. By operative reasons, each center receives only people listed to their own auspices. In 2006 there were 4 trusted centers in the County, which usually had a better quotation of doctor per 1000 inhabitants. It appears that these centers had higher availability and probably less referral proportion than the ones.

Some problems may appear: under-registered observations in some smaller groups (as academic hospital’s outpatient clinics) may lead to a larger variability in the results, influencing the final OR. Unmeasured risk factors for high participation in health care tend theoretically to accumulate, and may lead to overestimating the impact of the studied factors. Regarding some features as civil status or usual residence, it may be unclear whether variables as “unmarried”-“divorced” or “rural”-“urban” are registered and updated consequently or possibly include a partial migration across groups during the year studied.

In conclusion, the properties of the Swedish health care system open new possibilities for database research, unexplored until the date. A carefully maintained registry is particularly helpful in dealing with the possibility of loss of continuity and misidentification. The ecological model can be applied in a similar way as the original studies did, though the results are not comparable straight forward. The ecological model can not establish cause-effect relationship between variables and phenomena, but appears to be a good approach for assessing networks. Some new periodical updated evaluations across time may probably risen further interesting information for management, research, and policy making.

Slutsats

Socio-ekonomiska variabler i ett samhälle (dess egenskaper) har ett samband med hur vârdsystemet utnyttjas, men ingen orsak-effekt samband kan fastställas.
Den nya variabeln "patient-month" är ett bra verktyg för analysen av deltagande i sjukvârden.
Det finns möjligen en överandvändning av specialistkliniker i länet jämfört med USA:s, men andra metoder bör testas för ett mer konkret resultat.
I den studerade perioden verkar de distrikten tillskrivna privata förvaltninsformer ha en bättre tillgänglighet och mindre andel remitteringar, men tolkningar bör göras med försiktighet.
Databasstudier i Sverige är tillförlitliga och har en stor potential för utbildning och besluttagande. Databaskvaliteten bör undersökas med andra metoder.

Referenser

1.White KL, Williams TF, Greenberg BG. The ecology of medical care. New England Journal of Medicine 1961;265:885-92.
2.Green LA, Fryer GE, Jr., Yawn BP, Lanier D, Dovey SM. The ecology of medical care revisited.[see comment]. New England Journal of Medicine 2001;344(26):2021-
3.Pappas G, Queen S, Hadden W, Fisher G. The increasing disparity in mortality between socioeconomic groups in the United States, 1960 and 1986.[see comment][erratum appears in N Engl J Med 1993 Oct 7;329(15):1139]. New England Journal of Medicine 1993;329(2):103-9.
4.Montgomery LE, Kiely JL, Pappas G. The effects of poverty, race, and family structure on US children's health: data from the NHIS, 1978 through 1980 and 1989 through 1991. American Journal of Public Health 1996;86(10):1401-5.
5.Tugwell P, de Savigny D, Hawker G, Robinson V. Applying clinical epidemiological methods to health equity: the equity effectiveness loop.358-61, 2006 Feb 11.
6.Geiger HJ, Borchelt G. Racial and ethnic disparities in US health care.1674, 2003 Nov 15.
7.Flores G, Olson L, Tomany-Korman SC. Racial and ethnic disparities in early childhood health and health care.e183-93, 2005 Feb.
8.Roos LL, Walld R, Uhanova J, Bond R. Physician visits, hospitalizations, and socioeconomic status: ambulatory care sensitive conditions in a canadian setting.[see comment]. Health Services Research 2005;40(4):1167-85.
9.Bliss EB, Meyers DS, Phillips RL, Jr., Fryer GE, Dovey SM, Green LA. Variation in participation in health care settings associated with race and ethnicity.[see comment]. In: Journal of General Internal Medicine; 2004:931-6.
10.Roos NP, Black C, Roos LL, et al. Managing health services: how administrative data and population-based analyses can focus the agenda. Health Services Management Research 1998;11(1):49-67.
11.Fryer GE, Jr., Green LA, Dovey SM, Yawn BP, Phillips RL, Lanier D. Variation in the ecology of medical care.[see comment]. Annals of Family Medicine 2003;1(2):81-9.
12.Dovey SM, Green LA, Phillips RL, Fryer GE. The ecology of medical care for children in the United States: a new application of an old model reveals inequities that can be corrected. American Family Physician 2003;68(12):2310.
13.Dovey S, Weitzman M, Fryer G, et al. The ecology of medical care for children in the United States. Pediatrics 2003;111(5 Pt 1):1024-9.
14.Ahmad AW, Dag E, Kurt S. The influence of sociodemographic characteristics on health care utilisation in a Swedish municipality. Upsala Journal of Medical Sciences 2004;109(1):33-42.