Home
A user's guide to data collected in primary care in England
Contents
1. uo uapuadaq e salanb pue Ajisads 0 s y dx 10 Jang uBiu sauinbay S SS UNB NA 1802 18 91 W014 s seuuons nb uleygo ueg Bulguasesd uo uoewozu poop 9482 AJUAId UO pue X S pue e uo uoguuoJui s pn 3u ipiqioui pue Ayjenb e ep qo Bunoyuow snonunuo2 uanoid 10 saseqejep p epi BA 1802 MO elep ajq ssa39y ajdwes jeuolen saaijoeid 10 ajdwes 1 JJB1S WO ELEP ASIA181u sasodind Bulo uow 10 S H S lin aplAoud ueg Auryewyoueq l asn 104 990136 0 bunnq uo Buowe 331 321d 19 U1 ay Woda pue se le 9 UO uonp lLo ul Aewwns 0 pasn aq ueg salpnys jitueJ pue saizijenbaul pue SASN 011000 0 p sn s L nb 8 0IX8 J M 01 pasn aq ueg 119802 pue 10 SE up doin3 y m lq yedulo2 su Bu ns 5004 IS Buowe peaids 592136 0 09 woy slu ned 000 008 D3439A03 uone ndod p us lqnd pue p l ldulo9 niss oons 10 D10391 e salpnys yoJeasal ul ajediaiuied 0 123 Jad 000 13 pled 592128 4 e spiepue s eyep ln ll liN qual l s u31 2d 19470 10 a e u s 1d 1 5109 160 S831 9B1d 09 JO JOMJ N sn304 u ll S3u4H0M N yaieasay ae Mewnd 1001 HBAN10S U0119811X9 wass asea LSANDIIN xe u s niodxq pue M ng
2. u81J0 sow SUOI IPUO Jo Jaquinu e 104 paanpoud AjuQ S SS UNB AL 19A eu0 B 1 10 euol1eu 15 100 MOJ Je HUISSEN 10 5150 S A PUOnPu pue 024 n Rid 18 P AOId UB Buiquasaid 40 10 pue 1802 ay 10 p10991 l ld lo2 pue ajeinaae Ue S pIA01d Aa jod wozu pue Buiquaseid 1J0 lu0ul 0 suollesiueBJo 1lu 1 J ip Aq J 8AISu9 X8 D ST suonlpuo2 w413 Buoj Auew JO S10198 491 pue pue Bulyjows 6 a S10 28 ajAjsa4i 40 Buipue si pun 1818815 B N 0 Bl1u3 0d e 1S3 0IN Aq lqiss p U UM MO SS82328 JO S S09 e qo y pue l y A01duii djay 01 pIA01d ueg 1 UO 590 194 05 pue p q ow UO 290 S pIA01d su Bu ns saljuoyyne yyyjeay pue 3004 000 6 1noqe pue Bu3 ul s 317221d ej u 6 JJE UO ajqejiene UO JEUOJU e ldo d 01 10 6y 1noqe pue Bu3 uoiejndod alnu e uonpindod 09d uo qE 9p1SU02 s lueA S8211981d 810 Jo e 000 0SZ 0001001 09d 940 JJB3Id e pD34 9A03 uonejndod Je uol jiu 293 1noq ae S1902 9 s uou nv Bulag U011d1i19S31d e 6 Aq pasn BuI se 13u e sjafpng Buiquoseid Buas
3. e 0 unni3s oriqnd 1842316 NOJ B 0 99u8ulu0Ji d omawel4 590 021 pue 1 s 02d pue s n pid UO Uu0l Bul0Ju USI qnd e seqe1ep 140 pue SYND 995 19e11lu05 SIND Mau q p sejd y e I juauidojanap 10 sealy 8182 611050 10 Ayunwwos Jo Buiquoseid UO UO JEUOJU ON e aAg juaijed jEnpiaipul ou se e ep ul 0 ajqissod jou ajqeylene e ep parenge juQ e 9182 8211112 10 88920 0 UO UOTPULOl L ON UOIJEUOJU 82101 2 ON 9 pue swej Ajahie S SS UNB NA DEODHOM 49 UO uolBuuoJul S pIA01d e 9 1011395 s s 83113B1d jesauay ul A8AInS Aqpiquol qenuuy yuno ay l yed yoo Aue elep Aupiquow 09 ul p 52u u dx Je 91 01 1 00 181 599120 4 e suonpuo2 paljloads YUM 540 0 Buljuasaud ajdoad jo Jaquinu ay uo UOIJEUUOJU AJAWI S DIA01d 19A IEUOHEU pue e50 15 e ep SIND Mau SS 22B 0 PIPMIOJIYBIENS sabeyoed Jojeaipul 9183 1820 Ul pasn AjUOWWOD e su Bu ns 08 01 0 Woy S LA S8311381d JO J QWNN e ali auo Aue je e ep Bunnqiunuo3 59212840 Jo Jaquinu y uo Bulpuadap ajdoad 000 00 01 0007009 000 6 1noqe pue Bu3 ul s n pid jesauab JJE UO jqejrene UOHPULOL e ldo d uoliu 6y 1noqe pue 6u3 Jo uonpindod e uone ndod S8311981d 0 5
4. Contents eoeooooooooooo GLOSS ONY u 0 AREE EAER akawa dania 2 Executive summary with key messages and sources of primary care 3 Chapter 1 niiina adie 13 Chapter 2 Connecting for LUT Ra 15 Chapter 3 Using data from primary care to improve health 17 31 Needs assessment health service planning and COMMISSIONING 18 32 Regulation accountability and performance 19 3 3 Clinical governance and quality 21 SA MONITORING IEAM INEGUANM GS um uu adama Bs 21 39 Monitoring healthcare Uses c Sle ales tebe ieee 22 3 6 Monitoring death rates among patients of general practitioners 22 Jef PIESCHIDING POlICYANd PRICING uu u ages aa aaa eae 23 38 Pharmacosepidelmio 8gy u fetes creed L u puma uwa Aa 23 39 Resource allocation risk adjustment and case miX 5 23 310 Financial flows and payment by results 25 311 Public health research and health services aaa 25 Chapter 4 QOF data and OMAS
5. This report describes the main primary care data sources available in England and summarises some recent developments Comments or suggestions about the content can be sent to the main author Dr Shamini Gnani s gnani imperial ac uk Key messages are listed on pages 4 and 5 Sources of primary care data are tabulated on pages 6 to 12 A user s guide to data collected in primary care in England Key messages Chapter 1 Introduction e Data in general practice are population based and in many cases the patient record extends from birth and includes details of a patients diagnoses management and health outcomes e The introduction of a new GP contract with a Quality and Outcomes Framework requires general practices to routinely record detailed information on clinical management Chapter 2 Connecting for Health e Connecting for Health aims to centralise the electronic records of 50 million patients e The Health and Social Care Information Centre aims to combine information systems for health and social care Chapter 3 Using data from primary care to improve health There are several expected benefits from general practices improving information management and data quality These include e Improving patient care within the consultation high quality data about patients at the point of care ensures that there is legibility of medical notes data are complete and comprehensive there are alerts to clinical errors and appropriate war
6. Buru Aq 592138 0 p n f s ul Sjualjed le UO 211 00099 01205 pDV Z Xipuaddy Aanins B Ul 5 8 jeyidsoy pue saamas punwwo3 10 asn Bulquaseid uo UO JEUOJU BPN IU 0 D N e 51 Su0lRBlIBA 1U100029 0190S pue q5 uo UOIEWJOJU 0 dn s s G xipuaddy yuauidojanap 10 sealy 89920 0 MOJS e SI qlenb eeg e upu 0 3901380 10 ssauBul iM uo uapu d q e s nillqe pue Ayjenb 1134 pue s10181I 128J u D ns ui 8024 u0 1u pu d 8183 HBPUO29S Jo asn y UO uollBuu0Jul payu e Buiqii3s81d uo 40 ON e 26 L661 Ul INO Paley sayel 1 19002 d9 ul spu81 1olluou Ajayeinaae 0 JQLUN sassauyean Mo 8q 888 SANDIW S S NO SS822B Jep 10 S1802 ejep ql ru pue 211 0 03 0120S 3 udeJ6ouu p 19A JENpIAIpUI SUIe u09 e e ep 0 3129 Mewlid p005 e 540 UMA 9401161 19402 10 p 0281 l ld l02 e S pIA01d e 1jousdeus AyeaA e BuipiAcid ul pue sdnou6 abe Aq pue 54 Aq suone1 nsuo3 jo sayel Bululwexa Ul nJd H e su Bu ns 1 5004 Ile 10 1u3313d 06 s039d 9 Z SS019B peaids 590128 4 s jeM pue pue Bu3 ul 891 321d 09 wou ajdoad 000 Z0G 19A0 pue g sy npe 0069 noqe pue spjoyasnoy
7. because this function is not yet available in GP systems or MAS Data that are postcoded would enable linkage with census data Other identifiers such as NHS number may allow linkage with hospital admissions and mortality data This could lead to identifying those primary factors that lead to hospital admissions Furthermore PCOs would be able to produce additional practice and patient level reports for practices such as funnel plots Currently few PCOs are providing these additional reporting mechanisms but there are now several software solutions which aim to do this such as CHART from PRIMIS and A3 from Apollo systems Hence the use of QOF data by PCOs for such purposes is likely to increase Although the QOF attempts to measure the quality of health care delivered in general practice it may skew the focus of clinical care by its failure to cover all aspects of general practice especially those which are less readily measurable Limitations of QOF data There are a number of key limitations It is not possible to calculate age standardised prevalence rates for each of the chronic diseases in the contract framework This could be modified by linking information that practices hold on their patients age and sex Until recent incentives general practices did not systematically code the ethnicity of their patients This information is important in determining whether access is equitable and in contributing evidence as to whether certain d
8. information is collected as at 1 January 2005 from GP systems and a snapshot of all QOF disease registers collected as at 14 February 2005 National Prevalence Day to calculate disease prevalence QOF results are fed back to practices PCOs and strategic health authorities PCOs receive practice level data and strategic health authorities receive PCO level data PCOs and practices can access information at any time about their QOF achievement against their aspiration what practices thought that they would achieve the estimated relative prevalence and their current achievement payment From April 2005 GP practices have received financial payments that are based on their practice list size achievement data and aspiration data held on QMAS Although data are not interrogated by MAS GP software systems had to pass a data quality check The accuracy of QOF data is important as GP practices receive financial payments according to the quality of care they provide QOF data aims to give GP practices and PCOs objective evidence and feedback on the quality of care delivered to patients measured against national targets set out in the general medical services contract QMAS allows GP practices to analyse the data they collect about the range of services and the quality of care they deliver such as maternity services or chronic disease management clinics It is intended that GPs have a financial incentive to treat most patients in the community rather
9. the responsibility for the database was with the Medicines Control Agency MCA The GPRD contains information entered by GPs onto their practice computers the number of practices varies but is typically around 300 Data collection is available from 1987 for a limited number of practices Most of the general practices that participate in the GPRD provide data of sufficient quality for its use in research an analysis by the ONS showed that the 211 general practices passed all data quality checks during the period 1994 to 1998 inclusive Appendices There are several key strengths of GPRD data First it is population based the population is large enough to study rare diseases information on preventive care and secondary care are recorded and finally there is access to original records The GPRD has been mainly used for disease epidemiology pharmaco epidemiology and health services research A limitation to using the GPRD has been the cost in accessing data but a new licence paid for by the Medical Research Council will fund 50 projects to use the database free of charge each year for five years and will also pay for extra staff and a support service 30 4 National Database for Primary Care Groups and Trusts This database was developed by the National Primary Care Research and Development Centre in Manchester and the University of Manchester s Department of Geography as a resource for primary care researchers practitioners and managers The
10. 0006 uie ug ul spjoyasnoy ayeAud Jo ajdwes S Al 81U8S 810 1 PUONEN e p s ao uone ndod Jad udl Iw Z g 1noqe 5150 e saido 10 Ayauen e UO u0p pasnaoJ Ajjeaiul a JejnBay e 5892010 au 10 yed y3eqp pue juawissasse Ajijenb ezeq e SIIDIS Juawabheuew uolewojul pue qilenb eyep Ul S1018 11198 UOIEWOJU 1820 yBnoiyy 592112810 81e9np 0 SI WIE UIEN e SIININd s rd uonpuuo u a129 Mewnd J1 M0 qe1 pisu03 aq 0 ale 5 zead e jo S3802 l ld l02 0 spunod UOI IW 93 1100 1802 e sjualjed Jo sn1e s 211104029 01205 pue xas abe y 01 U01 8 81 1 90126 0 jesauab u aseasip 10 swayed y paulwexa ay e pd DSIN n biq fel u n ul Aaning Ayipiquoyy yuno 8211281d jesauab ul uo UOIJEUOJU JO DINOS UIEN e Aaning pjoyasnoy asoding pue 3inos gt x G 3 l 2 d 3 o o x W SIINIHd B A leuoneu p pu yx pino H M N ye s y 20111 0 34L 6 xipu qdy s s aAlsuajul 9 doad SS l 85920 0 ayew pue aul suods 0 uo do omyau pue BenBue pue sia aidalul SINOIW ayepdn o si riddns wayshs e 8 x pusddy aas juauidojanap 10 sealy e ep y eap 10 0 D N e W JS S
11. 27 5 Base Studies ineine ap attire alee 32 5 1 Primary prevention and public health surveillance 32 Case A How do we monitor the prevalence of adult and childhood obesity 32 Case B How do we monitor the prevalence Of smoking 7 2 34 5 2 Chronic disease MANAGEMENL ee ceccccsecsesessesseesesseseeseeseesssseesecseesuesesieeseeseestestsseeseesess 35 Case A How do we monitor diabetes Mellitus 35 Chapter 6 A surveillance system in primary care annassa 37 6 1 What is the role of a survey sesrsiisensiannisiin araa niaaa 39 6 2 What is the role of a register ses a 39 R l rant65 bo a 41 asay nayb 43 1 Routine NHS primary care activity data U U 43 2 RCGP Weekly RETUIIS 44 3 General Practice Research Database 2 2 44 4 National Database for Primary Care Groups and Trusts 22 45 5 General Household Survey GHS c cceccccseeccssssesseessssssessssssesessesssseesssnsstseteeesteetesceseeeesnees 45 6 Fourth Morbidity Survey in General Practice 4 46
12. 7 Primary Care Information Services 15 46 8 Morbidity Information Query and Export Syntax MIQUEST a aaa 47 9 Primary Care Research Networks Trent FOCUS 7 47 10 Practice based disease registers 48 11 Practice based health promotion 48 12 Prescribing Analysis and Cost PACT 49 13 Quality Prevalence and Indicator Database 0010 50 50 15 Secondary Uses Service SUS asesina 50 lb IERO s sen 51 17 Health Improvement Network THIN aan 51 B P TE E A TE sss 52 19 UK Clinical Research Collaboration Lu sada asandan sada adda a 52 Abhbrematinns c v c E E E AEE E E RD 53 Glossary Administrative data information that is a by product of administering care mainly for payment or reimbursement for services provided or to meet regulatory requirements It typically includes information such as patient demographics diagnostic codes and procedures performed Clinical data refers to the clinical attributes of patients and represent factors that health care professionals use for patients such as symptoms e g chest pain vital signs e g blood pressure and lab test results They are the types of observations writ
13. Ayipiq1oN asoding pue 3inos walshs v9d SHG aut yno T n jeuoljeu 18 ajqejeae Apeayje S L S BWI w YIN Apuana ajgeyiene ejep s4e 1 S A 520 pue 90i0eJd jeauab ze ezep s L s 10 A1HQE EAV e 5001101959 0 UO p8138 03 UOIPEWJOJU 01 e ep 3I3souBeIp uogippy e x s pue abe sjualjed uo UOIEWJOJU 01 e ep jo BBXUT Z s s Buljows pue Ayisaqo Jo 001139 02 UO p linb l S JOM 18u1InjJ e u01 8 o3 duaN jul Eli Jaded ally yeay 21 010 M N e L Xipuaddy aas 5004 10 suosnedwo3 moje 0 Seale 1u818 JJID Ul 59151 aJedwod 0 Aqijiqedea dojanaq e e ep say 10 y BulA01duii 10 549 01 SLyauaq 9 0 e U01181ID81338 ue pue e ep 40 s ul j pin6 dojanaq 0 Xipuaddy juauidojanap 10 sealy suoseal 69101 00 pue jelaueuly 891UU29 10 ajqiseajun aq Aew 9vd 0 5109 p8 s BBns y Jo Auey e S1 pi0SID 3IJ198ds 104 Buiqiu3s id uo uoneuuoJul 01 payul aq JOU ue9 e x s pue abe Aq Buiquosaud UO UOIJEWJOJU OU 99 sjualjed 0 paul aq 104 Ue e 9A 1520 e e paljdde ejep AaAins 60018 UO S B H e p9 99 09 A WUOJIUN 10N e 990126 0 uaamiaq e ep Aseunaae AY l SUOIELIEA DIAA aseesip yeay
14. among patients of general practitioners e Prescribing policy and pricing Executive summary e Pharmaco epidemiology e Resource allocation risk adjustment and case mix e Financial flows and payment by results e Public health and health services research Chapter 4 QOF data and QMAS The national MAS database currently holds indicator data for 8575 practices in England e QMAS is a national web based software tool developed for implementing the new GP contract e Data from practices are aggregated to maintain patient confidentiality and for every practice a set of quality QOF scores is calculated Information is collected on the number of patients with a particular chronic disease condition and on the care they receive There is potential to link QOF data with other data collected in primary care and with other data sources This would allow practices to be compared adjusting for differences in the underlying population to help understand differences in the quality of care provided QOF data have limitations and these include not being able to calculate age standardised prevalence rates adjust for socio economic and ethnicity differences in the population or adjust for inconsistencies in diagnosing and coding of disease conditions between healthcare professionals Chapter 5 Case studies e Obesity is predicted to have an effect on the health of the population equivalent to tobacco smoking e The Health Survey for Eng
15. aseasip 3Iu01u5 10 s ni 339u BA81d SD e S SB2 10 1090 16 29 e ep 10 Ayjeno S SS U B NA 9 Aq p sn Buise 13u e s on pid Bulynqijuod 0 10 y9eqp88J BWOS S pIA01d 9182 ul spu 81 Jo BuLO UOW 10 Salas awl pIA01d UB e Bupyewyouag l asn 10 599112 0 bunnq nuo Huowe UOHELEA 8311981d 1 1u BY Q 8183 10 spiepue1s JUALIND UO UOILEWJOJU 01 pasn aq UB e e ep 21U100029 0120 B L PU S UI0S S pIA01d e Buiqiasesd uo UO CUOJU poop e 8183 pue x s ae UO UOIJEWJOJU s pn 3u e uonpuuoyul qlpiqloli 31 duu03 e Joe uoneuuolu J0 40 yed se e ep JO uonealjqnd s s jeue fEUOHEU BL DUf e 1400005 pIA01d e aseasip lulu ilelo ds SUIEWOP PDIUI 9 S19A02 e i gnu piju03 1 npd saunsua ejep e AN ul S8211981d JO 1 e su Bu ns 599196 0 00G punole woy p l sib l Ajjuauna squared uoliu q Ajayewixosdde Sp10991 UIe U0J pue Bu3 ul s 3 15e1d Jg1 u B 0099 Ajayewixosdde S19A02 YOIYM P eD SYND pue 5901021 pue Alenp 01 ss8228 u pIM 0 51 WIY e ul do d uoliu 0G s on pid 0029 Ajayewixosdde 104 CLEP 62101 2 SD OH p 3i Ano03 uone ndoqd ealy Ayuoyny yyeaH 31637249S S19A02 p
16. be supplied as flat text files including data quality indicators to inform researchers of the completeness of data recording in practices Additional information will also include the geographical location of practices by health administrative area and socio economic indicators at individual patient level Practices are trained to record information and data collected from practices are subjected to a continuous programme of data quality control Recording is assessed against various quality indicators and national statistics are used for comparison THIN quality standards cover a wide area of clinical recording including asthma coronary heart disease diabetes mellitus epilepsy menopause hypertension hypothyroidism leg ulcers heart failure warfarin use lithium use oral contraceptive use pernicious anaemia rheumatoid arthritis secondary stroke prevention lower back pain mental health and smoking Practices will receive quarterly reports measuring performance in specific clinical areas 18 UK Biobank UK Biobank is a project currently under way which will follow the health of 500 000 volunteers aged 40 69 in the UK for up to 30 years Its aim is to build the worlds largest information resource for researchers to develop new and better ways of preventing diagnosing and treating common illnesses such as cancer heart disease diabetes and Alzheimer s disease It is funded by the Medical Research Council The Wellcome Trust the Dep
17. data in the near future Dr Simon Griffin Programme Leader Medical Research Council Epidemiology Unit Cambridge References a White K L NEJM 1961 265 885 892 b Scambler A et al JRCGP 1981 31 746 750 c Barker D BMJ 2003 327 1 428 1430 d Jick H et al BMJ 2000 321 1190 1195 A users guide to data collected In primary care in England Shamini Gnani and Azeem Majeed Department of Primary Care and Social Medicine Imperial College London Published by the Eastern Region Public Health Observatory on behalf of the Association of Public Health Observatories Acknowledgements We would like to extend our thanks to Dr Douglas Fleming from the Royal College of General Practitioners Weekly Returns Service and Mr Paul Bingham from the Eastern Region Public Health Observatory erpho for kindly commenting and reviewing the report and the team at erpho for their initial comments on the scope of the report Title Publisher Date of Publication ISBN Further copies from Copyright information A users guide to data collected in primary care in England Eastern Region Public Health Observatory erpho on behalf of the Association of Public Health Observatories January 2006 1 904389 09 0 www erpho org uk Quick Link 12899 email enquiries rdd phru cam ac uk or tel 44 0 1223 336101 Please contact the main author Dr Shamini Gnani s gnani imperial ac uk for permission to reproduce any part of this document
18. database links information on population characteristics health service provision and health status for all the PCOs in England Socio economic and demographic data derived from the 1991 and 2001 Censuses are linked to information on the characteristics and activities of general practice from the GMS statistics It provides a national tool for monitoring and evaluating the performance of PCOs over time Data are available to registered users from NHS organisations and the academic community At present the database does not include information on prescribing hospital referrals or hospital admissions data Once these are added the database will become a powerful tool for monitoring PCOs and in examining the links between population and patient characteristics and health outcomes Additional datasets are proposed and include the composition and organisation of PCO boards and their budgets information about the quality of care and prescribing information on measurable health outcomes such as morbidity mortality and health related behaviours such as smoking local authority data on community and social services health authority derived secondary health care information 5 General Household Survey The General Household Survey GHS is the main source of information on activity in general practice Respondents of the survey are asked if they have made contact with their GP in the previous two weeks the type and frequency of contact such as a ho
19. improvement Monitoring health inequalities Monitoring healthcare use Monitoring death rates among patients of general practitioners Prescribing policy and pricing Pharmaco epidemiology Resource allocation risk adjustment and case mix Financial flows and payment by results Public health and health services research 3 1 Needs assessment health service planning and commissioning Historically mortality data were used to estimate the burden of disease in populations This is due to the comprehensiveness and accuracy of collecting and recording deaths in the UK every death is reported to the Registrar General However it is also important to measure the morbidity associated with diseases not commonly recorded on death certificates due to its effect on both the use of health services and the wider costs to society Data derived from primary care can complement morbidity data derived from other sources such as hospital admissions data to describe the burden of disease in the community Unfortunately the population burden of chronic disease is not well measured and where it is available it is based on small studies There is a low degree of certainty of estimates of disease prevalence associated with single small studies when applied to the general population The small numbers of people associated with such studies means it is not possible to accurately calculate the prevalence rates of disease by age group or sex Furthermore sin
20. lt Imperial College erpho Ed London Eastern Region Public Health Observatory A users guide to data collected in primary care in England Shamini Gnani and Azeem Majeed Department of Primary Care and Social Medicine Imperial College London Published by the Eastern Region Public Health Observatory on behalf of the Association of Public Health Observatories ASSOCIATION OF PUBLIC SHIYOLVAEASEO HLIVA ASSOCIATION OF PUBLIC 9 98 HITv3H About the APHO The Association of Public Health Observatories was established in 2000 and has as a main focus facilitating collaborative working between the Public Health Observatories PHO in the UK and Ireland APHO was set up with the following aims To be a learning network for members and participants e To be a single point of contact for external partners e To be an advocate for users of public health information e 10 coordinate work across public health observatories Joint work is facilitated by e Each PHO taking the lead in a defined area to avoid duplication at regional and national levels e Acting as a major public health resource raising the public health profile at regional and national levels e Developing collaboration through links at regional national and international levels Further information about APHO the PHOs and their work can be obtained from www apho org uk Foreword Foreword As the late Kerr White and ot
21. maintain a high quality database of general practice derived data for use in medical research QResearch novv contains data from 468 general practices in the UK with records for 3 3 million current patients and 4 million past patients The database includes socio economic details such as patient s postcode But the data extracted will contain no strong patient identifiers Patients can opt out When it is fully established the OResearch database will be one of the largest aggregated databases containing records for nearly 8 million people The database will be open to researchers with ethical committee approval and information will be provided to answer their research question only not the whole dataset either at patient or practice level The costs of using the data will be carefully controlled to allow the scheme to be selffunding but allow good access to academic researchers Analyses will be undertaken to demonstrate the accuracy and completeness of the data and will be made available for morbidity analyses OResearch also has links with EMIS who supply clinical systems to just over half of all general practices There has been a recent joint initiative to provide information through development of a database on the number of people with flu By examining regional differences general practices and other parts of the health service will be aware of epidemics in time to introduce preventative measures 15 Secondary Uses Service The Seconda
22. pue Bulquaseid 0196 0 jesauab 10 MSU X D ST e S10198 1u02 eonn pulpud ulsinqullol JO Wass ay Jo JaNpold q e se padojanap A jelqiu e 19Vd 1802 pue sisAjeuy Buiqii3s id jesauab ul 0 uaye JOU sey PUOH DEL e ejep uonouioud yeay p seq n biq 1S3n0IN Bulsn yqiss op JOU 10 09d q lpnu paje 09 u q 104 ABU H UM SS833B 01 NIII 5 04 Auew Ag yse 104 pajeaoie Buipuni 612905 ou Ajjesauay e SJOMAWEI4 3IAJ9S jeualjeN 10 sjuawesinbas 188u 0 paysi qeysy e s19 sifai aseasip paseq aanoeg asoding pue aoinos gt x G 3 0 0 2 d 3 o x W 3I od pue ylp q 3I qnd 10 SN e SS 228 JO S1802 34 DD H pL 1lu uido A p 52050215 105 11050 se yans 5195 e ep JayjO YIM e S10128 SH S lpio261 pn 3u L sisAjeue pua JO MOJJE IM BWI p sajdey aag yuauidojanap 40 sealy Buiqiu9s 1d pue sisoubelp usamjaq u0 sul 12911PU e dnoig Mosiapy 31J11u915S e Aq jeaoidde anba syoaloud y e ejep Buisn 0 payseye S1802 WOS 31E e uon ldul09 Jeau nq juawdojanep Japun S 51 seqezeq e jeaosdde solid sauinbad 0 55922 e
23. quality of services The requirements of potential users of primary care data will differ Researchers may request person based datasets for independent analysis while clinicians and managers may request information already analysed Meanwhile the public may demand information that allows them to learn more about the range of services offered and to compare the quality of care among practices The potential of data collected as a by product of the new general medical services contract may help establish a system of surveillance in primary care However this is a complex area There is first a need to understand the limitations of GP contract data and to develop expertise especially when promoting its use in appraising health policy service developments and supporting public health activities In the United States there are more widespread analyses of administrative databases and surveys but with OMAS data which covers approximately 8500 practices there is great potential to improve the understanding of disease and illness Before data held by general practitioners on their computer systems can be used for surveillance two key issues need to be resolved First there is a need to improve the quality of electronic medical records in primary care particularly in the recording discipline of doctors Using databases such as GPRD it is now possible to get good information on chronic disease management at a national level However studies that ha
24. rates Referral rates can help predict the demand for specialist services and monitor changes over time For example the number of referrals is expected to increase as the population ages even if the prevalence of disease stays constant This suggests that some targets set may be difficult to achieve without either a large increase in the provision of specialist services or radical changes in the balance between primary and secondary care services 3 6 Monitoring death rates among patients of general practitioners Before the conviction of Dr Harold Shipman there was no national requirement to monitor death rates among patients of general practitioners A copy of the death registration of every resident who dies in a district or elsewhere is sent to the Director of Public Health District mortality data are processed by the ONS and are available as Vital Statistics VS returns VS returns contain some details of the number of people who die in the district the ICD code relating to the cause of death and data are presented by age and sex Post Shipman the system of death certification has changed 2 The Department of Health is working on the development of a single database containing information on every deceased NHS patient including the identity of the GP or GP practice with whom the patient was registered and the cause of death Monitoring mortality trends requires annual deaths to be examined because of the small numbers involved The data
25. to attend reviews and have been invited on at least three occasions in the preceding 12 months Chapter 4 e Inappropriate review e g because of frailty or terminal illness e Patients are on the maximum tolerated dose of a drug but are not within the requirement for a specific clinical indicator e g blood pressure control e Patients newly diagnosed within the practice or who have recently registered who should have measurements made within three months and delivery of clinical standards within nine months e g blood pressure or cholesterol measurements within target levels e Patients who decline or refuse drugs or treatment e Patients who are allergic or experience an adverse reaction to a particular drug or the drug is not indicated or is contraindicated e Where a patient has not tolerated medicine e Where a patient does not agree to investigation or treatment e Where a patient has a supervening condition which makes treatment of their chronic condition inappropriate e Where an investigative service or secondary care service is unavailable Potential uses of QOF data Using QOF data it is possible to describe differences in the quality of care across the clinical disease areas Interpreting this information may prove difficult It is possible to compare QOF scores by practice and by PCO QOF data could be used to calculate the crude prevalence of disease and thus to estimate population prevalence of disease across England This all
26. using QMAS data and other data sources The new GMS contract has for the first time a quality and outcomes framework 00F which has 76 performance indicators covering 10 chronic disease areas coronary heart disease and left ventricular dysfunction stroke or transient ischaemic attacks hypertension diabetes mellitus asthma chronic obstructive pulmonary disease epilepsy hypothyroidism cancer and mental health The framework also covers areas such as patient records and information patient communication education and training practice and medicines management patient experience and services such as child health surveillance and maternity and contraceptive services The national OMAS database currently holds indicator data for 8 575 practices not all practices in England have contributed data as involvement in the framework is voluntary Data are organised by practice and are aggregated to maintain patient confidentiality For every practice a set of QOF indicator scores are calculated The maximum score a practice can achieve is 1050 points The QOF scores achieved are then converted into a payment in accordance with the GMS contract Payments are then adjusted for practice list size and practice disease prevalence The National Health Applications and Infrastructure Services NHAIS or Exeter system administers the cancer screening call and recall system and patient registration NHAIS provides QMAS with information on list sizes
27. 180 AHU A Ld pue x s abe UO uoneuuoJul s pn 3u UOIJEWOJU A1DiQI0uu Bu pi0391 31 duu03 e su Bu ns Jang 8311381d jesauab 16 lQEHIPAR JOU SI UOHPULOL uol iw 1noqe pupibuz u sisni pue 50 aiea Aiewud V e 007 01 00 ua mz q WO Bl181119 olsni ul UO sanea papnyoul S8311981d JO J8quInN 59012610 10 uolsn 3ul 10 61 1 u01193 s uo B lp d p do d uoliu G E 0 e D3439A03 uone ndod pue Bunoyuow ul ASN 104 s 99d uo UONEWOJU u j seq 0 0 UHP H 10 1 94 1 0 aly Aq paysi qeys e sjsniy pue sdnoi9 ase Mewnd 10 aseqe eg IBUOnEN abieyo B BIA SJasn 19470 pue 5 Aq aujog Sey 1809 511 UO IW 2 Je u01 38 03 JP JO 5150 SUORELEA 80211981d 1 1u pue eale Ju gulle 1 pue gauajendid aseasip ul spuasy all Ba 5921 yqjeay 10 pasn A Buise813u S 29H apis Ajulew s lpns qpol5ololul pid oopulpud 100 Aveo 0 iledi ulid 100 e se dn jas E U 0949 seqeleq yoieasay n biq asoding pue ino 599128 0 0 y9eqp paseq q m ADEP HY e salpnys n t b 10 saljiwey Jo Buruj ajqeua 0 uoneolnu pi pjoyasnoy 10 asf Ains asn uay pue s poo sod
28. 4009 01205 J0 UOH DDV s s u uido j A p 0 se iV Ayepuogas Jo ddns y UO UO JEULIOJU ON e suoissiwpe jeyidsoy 10 sjeuajes jeyidsoy Bulquaseid uo UOIeWJOLUI Aue ule u09 18 104 s30q e pue dn jas 0 tin uoj e aye ued 5120 04 uo sjui 1934Ipui pue sisoubelp UdIMJAq 19891ID ON e 3004 SE yons sesi 9 104 UO JEWWOJU pIA01d 0 pasn aq 100 Ue e 10 3u lu dx hli aAey suayoseasal 124 MAU e SEH e aseqejep y 01 ssagae SNOJS SIU dn019 MosiApy poli pue 31J1u815s e Aq jeaoidde uinb i syoaloud y e asn 0 sn dx e3 uu38 qo uBiu seuinbay aseqejep jeaijsizeys ugy 1941 1001 yoweasal e se dn 195 Ajjeliu S SS UNE M suasn p8181s B 1 10 aseqejep y 0 SSIM lte nB i pa epdr pue Bu3 u 1 99d 2 uo S ND pue 210 0409 0190S uone ndod S pIA01d 1e3u1 e u ul spu n Buliolluouu 10 e ep S H S lil UB Dubueuluousq ul sn JOY 999110640 bunnq suo Buowe UOHELPA 8311981d 131ul y pue 8163 spiepue s 1u din UO uoleuloJul aplAoid 0 pasn aq ue e og 9 AIJ 10 1594 yoes abieyo 3914 aseqejep ay asn 0 si lold QG Moje IIVM i39uno9 Yyoeasay Jepa ayy Aq 10 pled aouaal Mau e pue yojeasal 101 pasn J pIAA e Buiquaseid UO UOIJEUOJU p009 9
29. 906 Spuy OMY UNI 0 1594 Jad 00010073 Noe 5150 saseasip je Jo 9u eA8 d pue u pioul AD M ay uo 51 009 jenuue s 3np01d e Buisn IIAM saseasip qe3 unuutu03 pue 1059 p 2 l s 10 adUapIIU ay s usi qnd e m s suinjay ADI JA 4924 Ajjeuoizeu pue B30 NO 918 SS822B BTEP 10 S S09 e 8024 pue s nuou ne Ule u 40 uoun aAljesiulupe 0 jesHayul se mse ul 01 1 n3 JJip g ep Bul133 03 10 5150 5024 pue saaioesd jg1 gu B Buowe sduewoped 0 D ST 890136 0 jesauab 01 pled S 2UBAQ E pue 598 10 tu s s ay Buajsiulwpe ufnolu p3139 07 ep SIND s m jesauay Mewnd SHN unnoq asoding pue ino A Aaevuurid Jo s Sinos gt x G 3 0 2 3 o x ul elde Jad spaq jeqidsoy 10 S S E 2 dS se yans sajqeuen Ajddns UOR DDV e Saw pue sisil BulIBAA UO SB AA SE 312I 2110504 asn pue Bulquaseid UO UOIJEUUOJU JO UOIIPPY e p Xipuaddy s s saseqeyep aseo Mewud 139470 wouy uonpunolul Bulsn Aq pajuawa dwod aq p n09 e U AL D 104 asn p3se813u e Sp40291 9 pue s1 u1oui Huu 10 poyjaw e Jo JuaWdojaAaq s p031sod 4910 Aq 19 2 sapooysod 831 381d Jo asn ayy q Ayjeniul Uonp lo l 21 0
30. B Weiner JP Use of risk adjustment in budget setting and performance measurement in primary care Paper Il Advantages disadvantages practicalities BMJ 2001 323 604 607 24 Controller and Auditor General Tackling Obesity in England London National Audit Office 2001 25 Department of Health Hea th Survey for England London Department of Health 2000 26 National Diet and Nutrition Survey accessed by http www food gov uk science 101717 ndnsdocuments 27 Office for National Statistics Key Health Statistics from General Practice 1998 Office for National Statistics London 2000 28 Gray J Orr D Majeed A Use of Read codes in diabetes management in a south London primary care group implications for establishing disease registers BMJ 2003 326 1130 1134 29 Hammersley V Hippisley Cox J Wlison A Pringle M A comparison of research general practices and their patients with other practices a cross sectional survey in Trent Br J Gen Pract 2002 52 463 468 30 Kmietowicz Z UK academics gain free access to database of patients BMJ 2005 331 924 31 The National Clinical Audit Support Programmess National Diabetes Audit is described at URL http wwwi icservices nhs uk ncasp pages audit_topics diabetes default new asp Appendices Appendices eoceooooooooo 1 Routine NHS primary care activity data 2 Royal College of General Practitioners Weekly Returns Service 3 General Practice Rese
31. NHS Executive General Medical Services GMS contract 2004 London NHS Executive 2004 2 Department of Health Choosing Health making healthier choices easier London Department of Health 2004 3 NHS Executive nformation for health An information strategy for the modern NHS 1998 2005 Leeds NHS Executive 1998 4 NHS National Programme for Information Technology Making IT happen Information about the national programme for IT Birmingham NHS Information Authority 2004 5 Department of Health Delivering 21st Century IT support for the NHS National specification for Integrated Care Records Service consultation document London Department of Health 2002 6 Connecting for Health previously NHS National Programme for Information Technology http www connectingforhealth nhs uk 7 Wanless D Securing our future health taking a long term view Final report London HM Treasury 2002 8 Department of Health The NHS Plan A plan for investment A plan for reform London Department of Health 2000 9 Department of Health Delivering 21st Century IT support for the NHS National strategic programme London Department of Health 2002 10 Home Office Freedom of Information Act 2000 London Cabinet Office 1999 11 Data Protection Act 1998 London Stationery Office 1998 12 Department of Health Reforming NHS Financial Flows introducing payment by results London Department of Health 2002 13 Department of Healt
32. QUEST compatible before receiving accreditation However there remains in many areas a lack of expertise in writing MIQUEST queries The Department of Health introduced several initiatives to overcome some of the problems with data quality in primary care These include PRIMIS a no charge support service designed to help primary care organisations improve patient care through the use of their computer systems and Health Information Project for Coronary Heart Disease a project to help practices and PCOs measure their progress in meeting the national service framework for coronary heart disease standards There are also independently funded data quality projects such as Primary Care Data Quality programme This is an educational intervention that aims to improve data quality by building on existing skills and knowledge within primary care of data recording and the use of technology A significant development in the improvement of coding is likely as a result of the introduction of the QOF The second issue that requires resolution in setting up a system of primary care surveillance is to reduce the burden of data collected by practices Decisions need to be made as to what data should be collected in a primary care data system This would minimise the development of adhoc databases A further key approach is to link existing data through data linkage systems Attempts have been made to do this by the Department of Health with the National Clinical Aud
33. These areas need to be considered when using survey data for carrying out surveillance 6 2 What is the role of a register Registration systems can be time intensive and expensive hence it is important to consider the need for a register A key component of the QOF and national service frameworks are that practices are required to maintain registers across many disease areas Typically a register in primary care contains personal details such as the patient s address date of birth and sex and disease status details of treatment and outcome and can be viewed by disease or risk factor status or by treatment Data collected by registries in primary care often omit socio economic and ethnic data information Variations in how diagnosis and classification of disease occurs can give rise to problems when comparing data between different countries between different areas in a country or between the same populations over time It is important to validate the information recorded and continually evaluate its quality as poor quality registers are of little use and may be misleading But comparing data derived from practice registers even within the same PCO has been extremely difficult In many cases the methods used to produce such registers including case definitions were left to individual practices and varied widely Interpreting the prevalence of chronic disease and treatment rates locally has been complex due to the variability in the com
34. about which patients on the register have died over time those patients that are newly diagnosed or whether they were admitted to hospital The usefulness comes with linking data across more than one registry Medical record linkage is the process whereby health records from two or more different sources and containing different types of information are brought together to provide a single file for the individual For example linking cancer registration statistics with mortality data enables survival rates to be compared between different groups of people with different types of cancer A registry must establish systems to maintain the reliable notification or identification of cases within the studied population ensure comparability and strict rules in diagnosis A revision must minimise under coverage cases not being included when they should be and ensure that duplication of cases within the register does not exist by keeping the register updated removing those who have recovered died or moved out of the area To date creating primary care disease registers by searching on coded patients with a disease and drugs was a relatively simple strategy but it is resource intensive It is likely that registers need to be coded prospectively using hospital discharge summaries and outpatient letters There will be a need to factor in the cost of coders at a practice level References References eeoceooooooo 1
35. advice and referral to specialist stop smoking service may help develop appropriate services to the need of the population However there are key issues that should be solved before developing a surveillance system for smoking First it is important to link the varied sources of data collection on smoking thus avoiding the numerous stand alone databases Second to improve the extraction of smoking data from GP clinical systems Third to take into account in analysing the data methodological issues associated with recording smoking status over time and the problem of repeated measures 5 2 Chronic disease management Case A How do we monitor diabetes mellitus The prevalence of diabetes mellitus will increase significantly by 2010 Diabetes mellitus is a considerable health problem for the individual and society The prevalence of disease is higher in areas of socio economic deprivation and among certain ethnic groups such as South Asians There is strong evidence that good quality diabetic care improves outcomes such as the prevention of blindness and heart disease Before 1970 most patients with diabetes were cared for in hospital but since then there has been a gradual shift in care of patients to primary care In this case history we look at the monitoring of the key factors that influence the health of patients with diabetes Evidence centres on good blood pressure control and blood glucose monitoring Many patients are not diagnosed fo
36. al practices Indicators were rarely made available to the public In many PCOs this information was combined with other information from age sex registers and published as primary care indicator packages Some of these data are now available to primary care organisations in the national database stored at the National Primary Care Research and Development Centre in Manchester 3 Most recently the NHS now collects and publishes information on general practices collected as part of the new GP contract described in Chapter 4 of this report Royal College of General Practitioners Weekly Returns Service The Birmingham Research Unit of the Royal College of General Practitioners RCGP funded by the Department of Health is responsible for the Weekly Returns Service which has operated continuously since 1964 and is now a fully automated system The Weekly Returns Service can provide weekly information on the total rates of consultation with GPs by age and sex as well as weekly consultation rates for any disease The system facilitates the estimation of annual consultation rates and since 2001 has also provided annual prevalence data For this purpose prevalence is determined by the need to consult for a particular problem or group of diseases classified according to the structure of the International Classification of Diseases during the course of the year The research unit manages a network of about 80 to 90 general practices in Englan
37. als to access their records By 2010 the NHS Care Records Service NCRS or data spine would provide a cradle to grave NHS record for each patient holding essential information that is accessible by different local NHS organisations 5 Connecting for Health in England awarded the national contract to British Telecommunications to develop software and to support the connection of all NHS organisations including GPs acute trusts and community and mental health trusts in a single secure system Connecting for Health has been divided into five regional areas consisting of between five to seven strategic health authority areas Southern London North East East West Midlands and North West Each regional area has contracted with a Local Service Provider LSP which is a commercial organisation that has bid for a contract to provide information systems to the local health community A new Health and Social Care Information Centre was set up in March 2005 based in Leeds 6 The centre is a Special Health Authority incorporating the Prescribing Support Unit It aims to combine the information systems for health and social care provide national leadership for data and information and respond to the information requirements of those in the NHS The first Wanless report the NHS Plan 8 and Delivering 21st Century T support for the NHS 9 all recommended that information technology in the NHS should support the development of electronic records Howe
38. and PCOs have mainly used this data to help set and monitor the prescribing budgets of practices Researchers have used PACT data to investigate variations and trends in prescribing costs for example between fundholding and non fundholding practices But PACT data provide very limited information on prescribing patterns for specific conditions At present the main method of monitoring prescribing for specific conditions is to use data collected from primary care databases such as the General Practice Research Database In the future data collected as part of QOF may help to link prescribing data with disease data 3 8 Pharmaco epidemiology Research in pharmaco epidemiology has important implications for public health and patient safety both in the United Kingdom and elsewhere The use of primary care data is powerful in examining the side effects and complications associated with both new and established drugs Data collected in primary care over time allow researchers to examine the long term complications linked with drug prescribing For example research using primary care data shows an increased risk of thrombo embolic disease among women taking third generation combined contraceptive pills However there are limitations to these studies due to bias and confounding with the potential for researchers to draw erroneous conclusions about the safety and effectiveness of drugs These observational studies can be carried out more readily compared to
39. arch Database GPRD 4 National Database for Primary Care Groups and Trusts 5 General Household Survey GHS 6 Fourth Morbidity Survey in General Practice MSGP4 7 Primary Care Information Services PRIMIS 8 Morbidity Information Query and Export Syntax MIQUEST 9 Primary Care Research Networks Trent Focus 10 Practice based disease registers 11 Practice based health promotion data 12 Prescribing Analysis and Cost PACT data 13 Quality Prevalence and Indicator Database PID 14 Research 15 Secondary Uses Service SUS 16 MEM 17 The Health Improvement Network THIN 18 UK Biobank 19 UK Clinical Research Collaboration 1 Routine NHS primary care activity data Health authorities and PCOs collect some information on NHS activity in primary care This mainly concerns details of patients registered with GPs preventative health screening such as cervical screening and childhood immunisation and item of service target payments for GPs for example to undertake minor surgery Information is also collected on staff employed in general practice such as GPs and practice nurses as part of General Medical Service statistics In the 1990s many health authorities began to use routine NHS activity information to produce performance indicators on general practices in their area These indicators were generally used internally and when they were supplied to general practices did not reveal the identity of individu
40. arch groups and networks The use of MIQUEST is part of the free training that PRIMIS provides to information facilitators who are employed by PCOs to help practices make the best use of their clinical computer systems The facilitators will usually run the first MIQUEST data extraction in each practice and then train a practice staff member to run subsequent extractions themselves However many practices have been reluctant to allocate staff time to MIQUEST training and data extraction 9 Primary Care Research Networks Primary care research networks consist of professionals interested in undertaking research in primary care Most networks aim to support new researchers help with project management assist researchers in writing articles for publication and fund more established researchers The UK Federation of Primary Care Research Organisations was established in 1998 to bring together all primary care research networks in the UK The aim of the federation is to promote learning across networks and to secure the long term future of networks Any primary care research network within the UK that supports the aim of the federation may join An example of a primary care research network is Trent Focus Trent Focus Trent Focus is a Collaborative Research Network set up in 1994 funded by Trent Regional Health Authority with the aim to improve research knowledge and skills in primary care A key objective of Trent Focus was to establish a networ
41. artment of Health and the Scottish Executive The project will provide information about volunteer s lifestyle such as exercise patterns and diet employment history and their medical and family history along with samples of their blood and urine Simple measures such as weight and blood pressure will also be recorded People are randomly selected via health registers and asked if they are willing to participate This information will be linked with the participants medical records so that researchers can study the links between the participants genes lifestyle environment and the diseases and conditions they may develop 19 UK Clinical Research Collaboration The UK Clinical Research Collaboration UKCRC is a partnership of organisations with the aim to establish the UK as world leader in clinical research The partnership is represented by the main funding bodies for clinical research in the UK academic medicine the NHS regulatory bodies industry and patients See gniting our potential An introduction to the UK clinical research network available at URL http Awww ntrac org uk News Summary UKCRC 20Introduction 20leaflet 20Dec 04 pdf The NHS Confederation is setting up a health services research network which has been supported by the Department of Health the NHS Service Delivery and Organisation Research and Development Programme the Health Foundation and the Nuffield Trust All these clinical research collaborations will requ
42. ates of cardiovascular disease The Health Survey for England HSE is the main source of routine data on the prevalence of obesity in England 25 The survey is carried out each year Each participant has their height weight and waist and hip measurements recorded However information on obesity at a local level is not available because of the sample size used in the HSE The National Diet and Nutrition Survey is another source of data on obesity that includes BMI and waist to hip ratios 26 This was carried out in 1986 87 and in 2000 2001 But similar to HSE it is not possible to undertake analyses at a local level Some health authorities undertake local surveys of obesity but these tend to be ad hoc and limited to subsets of the population and the results cannot be assumed to be representative of other areas of the UK In England the levels of obesity and overweight among boys and girls shows a steady increase from 1995 to 2002 Obesity in boys almost doubled rising from 2 9 in 1995 to 5 7 in 2002 The trends observed are more marked in children from households in manual social classes compared with non manual classes It is difficult to define obesity and overweight among children Currently children are defined as being overweight at the 85th percentile and obese at the 95th percentile according to UK 1990 reference curves This approach assumes that the prevalence of obesity is 5 and that of being overweight is 10 But there is little evi
43. base is to be a surveillance system that identifies GPs with high rates of mortality among their patients and thus allows GPs to be scrutinised more closely For effective monitoring of mortality rates at individual GP level records need to be linked between an individual patient registered with a practice and a named GP with that practice However the patient lists that GPs and PCOs hold often have errors Patients frequently change address without notifying their GP and sometimes patients on the list may no longer be present in the area This creates list variation a difference between official population estimates and estimates obtained from GP lists The difference is particularly large in inner city areas and may lead to inaccurate denominators and hence inaccurate rates 3 7 Prescribing policy and pricing In England the NHS community prescribing or drugs bill is about 7 billion per year greater than the costs of all other aspects of general practice combined Ensuring the NHS obtains value for money in community prescribing is an important task for the Department of Health and PCOs The main source of prescribing data is PACT data The collection of PACT data was initially a by product of reimbursing pharmaceutical contractors Now it is widely used to monitor the cost and volume of community prescribing in the NHS the implementation of prescribing policies and to ensure cost effective prescribing in primary care Health authorities
44. because GPs can exclude patients with the use of exception codes Practices may have an incentive to use exception reporting inappropriately to maximise their income by selectively identifying patients High levels of exception coding may be associated with practices delivering poor quality care or may be associated with areas of deprivation OMAS data are extracted from individual practices and aggregated Therefore there is potential for linking these data to other data sources at practice or PCO level This would confer significant benefits For example it would provide a much better indication of the effects of practice resources or population factors such as deprivation or ethnicity on practice performance It would allow benchmarking and inter practice comparisons and validation of community profiles an opportunity to develop a predictive model of expected prevalence by age sex and ethnicity at a practice level identify outlying practices This information would help contribute to the planning and commissioning of services and to improving the management of chronic disease by the primary care team and thus reducing hospital admission rates Population based data on lifestyle and risk factors such as prevalence of smoking high blood pressure and obesity could also be recorded Analysing anonymised patient specific data would allow much more powerful analyses of patients at high risk of coronary heart disease e g absolute 10 year risk gt 30
45. but may also require drug therapy Many of the indicators are measures of the process of diabetic care such as recording whether screening for eye or circulatory disease problems has taken place Other indicators have been linked to poorer health outcome such as identifying patients with protein in their urine who are more at risk of heart and renal disease There is a trend towards patients with diabetes being managed outside hospital in primary care settings especially with the development of GPs with special interest in diabetes Data from QOF can help decision makers when commissioning services for patients with diabetes Now there are diabetes prevalence data by general practice and by PCOs At present 00F data only provide crude prevalence figures which cannot be analysed further by age sex or ethnicity However if at a local level practice population data can be linked to QOF data this would provide more detailed information on the burden of diabetes and population need Further analysis on the type of diabetes and method of treatment such as diet controlled oral hypoglycaemics or insulin therapy would help determine where diabetic care would be most appropriate to take place such as hospital or within practice Disease severity among patients is important and can be determined by examining indicators such as glycosylated haemoglobin HbATc or examining prescribing data to identify patient with co morbidities such as heart and kidney disea
46. ct on health it is important to compare national survey data with information collected at a local level for example the registered population of PCOs who are recorded as obese Case B How do we monitor the prevalence of smoking Key messages Smoking is the single most modifiable risk factor for ill health In primary care the recording of smoking status on general practice computers is variable Smoking data are recorded on separate databases either within hospitals or PCOs and are rarely linked It is important to link the varied sources of data on smoking improve the extraction of smoking data from GP systems and to take into account in analysis the recording of smoking status over time Smoking is the single most important modifiable risk factor for ill health To date information collected on the prevalence of smoking is taken mainly from the Health Survey for England These data are then often applied to local populations to determine the local prevalence of smoking Some areas have undertaken their own surveys among groups such as pregnant women or among ethnic minority groups Current smoking status was obtained routinely as part of the specific socio economic enquiry used by MSGP4 and was used as a standardisation criterion for several of the analyses of disease prevalence All PCOs are required to collect smoking data In England each PCO has a stop smoking service and its performance is monitored by the Department
47. d and Wales Each week data are sent electronically from practices to the research unit The data sent provides the number of episodes of care and information on practice populations An episode of care represents one consultation between a patient and their general practitioner and is recorded as first new or ongoing The mean weekly consultation rates per 100 000 population are calculated based on first and new episodes The research unit also publishes mean weekly incidence for selected respiratory and communicable diseases and several symptom complexes including flu like illness compared to a background over the previous 10 years Other published statistics include information on chronic diseases such as eczema depression and heart failure The Weekly Returns Service could contribute to public health surveillance by providing information on illnesses and symptom complexes that may mark bio terrorist activity General Practice Research Database During the late 1980s VAMP Health started to install computer systems and practice management software in GP surgeries throughout the UK The aim was that data collected on these computers could be used for practice administration and research The General Practice Research Database GPRD owned by VAMP Health was taken over by Reuters Plc and subsequently given to the Department of Health The Office for National Statistics ONS operated the database on behalf of the DH between 1994 and 1999 From 1999
48. dence that these cut off points relate to morbidity or health outcomes The government has set a national target by 2010 to stop the annual increase in obesity among children under 11 years However in setting this target the government has exposed the lack of information on the extent of childhood obesity as children are not weighed routinely Undertaking physical assessments of children routinely is viewed as a poor screening procedure due to the balance of benefit and costs of undertaking an assessment But in view of the rise in obesity levels among younger people it may be justified to measure height and weight at school entry The benefit may also extend to measuring obesity at age 10 11 years and 15 16 years as policy makers may need to develop appropriate interventions at different points in time How should obesity be monitored for adults The majority of BMI recording in general practice is ad hoc Practices are most likely to record body weight when a patient first registers with a practice as part of a health check Weight may also be recorded by general practitioners or nurses when a patient expresses a wish to reduce their weight or a patient is taking a medication that requires weight monitoring such as the combined oral contraceptive or hormone replacement therapy or they appear overweight and are considered to be at risk of for example heart disease One method is to collect obesity data routinely as called for in Choosing H
49. e management Previously the NHS principally health authorities and PCOs monitored the performance of general practices and primary care using financial administrative data sources such as general medical services contract and PACT data The performance indicators they used were largely based on routine data sources of population health measures applied to general practice This was partly due to the variability of data collection in primary care and ad hoc information systems among general practices These primary care indicators were criticised for being crude measures of performance Existing performance measures for primary care trusts may mask the significant variations in performance that exist among general practices However using indicators to monitor performance at a general practice level requires reliable data collection and validation and this in turn requires good information technology systems which has implications on resources Data items from the QOF may lead to the development of more robust indicators to monitor the quality of care provided by general practices and PCOs Indicators based on routine data are easiest to produce compared with carrying out patient or practice surveys In contrast to primary care NHS hospitals have a more developed procedure for collecting information patient administration systems and Hospital Episode Statistics The collection of a minimum data set by each hospital allows the production
50. e which observes the same recording discipline Before the survey started doctors and staff from each practice attended three two day training sessions on how to record morbidity data Practices then collected data for two to four weeks before the start of the survey These data were analysed and any errors or inconsistencies reported back to the practices Once the morbidity survey started general practitioners and nurses recorded information on all face to face contacts with patients Each reason for consulting and the place of contact was directly entered into patient records on the practice computer Every consultation was assigned an ICD 9 International Classification of Disease Ninth Revision code When patients presented with more than one problem doctors were asked to record a separate ICD 9 code for each problem Although the number of diagnoses recorded was greater than the total number of contacts with general practitioners the vast majority of contacts were for one problem only Data supplied by the practices were subject to regular checks to ensure Its validity The practices that took part in the survey compared with the average general practice were bigger were more likely to be computerised and to show a greater interest in the collection of morbidity data However the sample of patients was representative of the population of England and Wales for characteristics such as age sex and social class The key strength of MSGP4 is t
51. e and Indicator Database QPID will improve the potential of primary care data to examine the prevalence of disease locally This in turn should improve the planning of health services and the allocation of resources according to where there is greatest need Data from primary care are useful in implementing and monitoring national service frameworks For example a recent study estimated the workload implications of the national service frameworks on coronary heart disease for general practitioners By combining disease prevalence derived from primary care with population projections it is possible to estimate the number of people with chronic disease locally and plan interventions to improve the responsiveness of health services Primary care data can be used to help plan and monitor the provision of services according to need and identify areas in primary care where services or interventions are effective or missing PCOs can use this information to help produce Local Area Profiles and secure better services by implementing evidence based care or best practice allocating resources appropriately and in investing in appropriate interventions including education and training The analysis of estimates of disease prevalence from the QOF may support local surveillance of chronic diseases and the monitoring by public health organisations and integration of epidemiology in service planning and delivery 3 2 Regulation accountability and performanc
52. ealth An option here may be to require general practices to record the BMI of all their registered patients and identify those with a BMI that is over 25 and 30 or measure patients waist hip ratio in the last 15 months Another option is to carry out local surveys each year The Scottish Intercollegiate Guidelines Network recommends that doctors should opportunistically take BMI and waist measurements at least every three years There is little comprehensive evidence on the effectiveness of strategies especially in primary care that reduce levels of obesity GPs may be influential in getting patients to change their diets but most advice given to patients tends to be disease specific The extent to which GPs provide nutrition education to patients is reported to depend on their perception of their own ability to influence lifestyle and confidence in their ability to advise patients about their diet The British Nutrition Foundation Task Force on obesity identified at risk groups as obese children and children with obese parents rapid weight gainers greater than 5kg in 5 years post obese pregnant women smoking quitters physically inactive and certain ethnic groups This may help target interventions either practice or hospital based in the management of obesity At present information on obesity is mainly available at a national level due to present systems of data collection To monitor levels of obesity more effectively and its effe
53. er work in monitoring health inequalities may be helped by local incentive schemes for GPs to record socio economic and ethnicity data on their patients 3 5 Monitoring healthcare use The monitoring of health service use has focussed mainly on monitoring hospitals and hospital admissions The use of primary care services was monitored last by the Fourth Survey of Morbidity in General Practice MSGP4 in 1991 1992 20 The survey provides information on the range of conditions presenting to GPs and their workload This information has been the basis of contract negotiations about workload between GPs and the Department of Health The RCGP s Weekly Returns Service provides more up to date information but person linked socio economic analyses are not available and do not include MSGP4 data The main method of monitoring prescribing for specific conditions is from data collected directly from primary care for example primary care databases PRIMIS or local disease registers PACT data the other main source of prescribing data allows the monitoring of cost and volume of community prescribing in the NHS PACT data are accurate but provide limited information on prescribing patterns for specific conditions The referral rates for specialist care are an important area for monitoring health service use the cost of patient care rises substantially on referral to hospital The NHS plan requires PCOs to have methods of monitoring general practice referral
54. esults from the QOF have been published allowing public access to information on the quality of care provided by general practices Furthermore following the conviction of Dr Harold Shipman there is a requirement for improved accountability of doctors including recording of mortality rates among patients of individual GPs 15 There is debate as to whether patients will choose to register with practices that perform better based on the indicator results published For most patients primary health care teams are their first point of contact with the NHS The care patients receive within primary care has a major knock on effect on the use of other NHS services for example on prescribing investigations referrals and hospital admissions Patients perceptions of the quality of care that the NHS provides are to some extent also determined by their experience of primary care 3 3 Clinical governance and quality improvement Many practices use data from their computerised medical records for clinical audit and clinical governance Some PCOs have found using data in this way as a powerful incentive for improving the completeness and accuracy of data recording and consequently the quality of care To be useful for quality measurement purposes clinical data must be computer readable Before practices became computerised GPs recorded most if not all clinical information on paper This meant that important clinical data in one part of the system was n
55. eveloping a process of quality assurance for data entered onto clinical systems by GPs PCOs are required to visit each practice annually part of the purpose being to audit the data collection processes The research value of using MAS and OPID data would be increased significantly if they were available at patient level Long term gains may also be achieved by training practices in coding and classification of conditions The monitoring of data by PCOs and feedback to practices may help in this Finally the identification of exception codes used by practices is not possible with the 2004 2005 data The ability to do this with 2005 2006 data may help to explain variations Hospital Episode Statistics HES data contains all records of inpatient care provided by NHS hospitals in England and when it was first introduced was initially thought to be of low quality Now it is being used increasingly for financial management and clinical audit Similarly it is likely that the quality and accuracy of QOF data will be poor at the start but will improve over time Thus it will be more useful for undertaking health services research especially in investigating primary care services Furthermore the linkage of QOF data with HES data may improve the analytical power of studies examining the quality of health service care For example whether there is an association between the quality of care of patients with asthma and the number of asthma hospital admissio
56. facilitate important and much needed observational studies of the natural history of a range of symptoms The opportunities for using information held in electronic health records in general practice are only likely to increase for several reasons The 6 3 billion investment in Connecting for Health formerly the National Programme for Information Technology in the NHS should start to realise some of its potential before 2010 The UK Biobank the UK Clinical Research Network and the topic specific research networks will all be heavily dependent on electronic data held or collected in primary care General practices now record information on clinical management which is linked to payments as part of the Quality and Outcomes Framework of the new GP contract This should lead to improvements in data quality and has renewed interest in the use of general practice data to estimate the frequency natural history and determinants of different diseases This succinct clearly presented Association of Public Health Observatories report written by authors with considerable expertise is very timely It brings together and summarises information about known sources of primary care data focusing on general practice teams rather than the professions allied to medicine the methods and systems used to access primary care data at the moment and sets out their range strengths and weaknesses It should be required reading for anyone who plans to use primary care
57. for paying trusts It also aims to support patient choice and diversity and encourage shorter hospital waiting times Payment will be linked to activity and adjusted for casemix Traditionally funding for hospitals relied on historic budgets and the negotiating skills of hospital managers Under the reforms to NHS Financial Flows instead of being commissioned through block agreements as previously hospitals and other providers will be paid for the activity that they undertake so PCOs will commission the volume of activity required to deliver service priorities adjusted for casemix i e the mix of types of patients and or treatment episodes from a plurality of providers on the basis of a standard national price tariff adjusted for regional variation in wages and other costs of service delivery 3 11 Public health research and health services research In Securing Good Health for the Whole Population Derek Wanless was critical of the lack of evidence of interventions that would lead to improved health Primary care data have been used in health services research typically studies have examined disease prevalence and treatment effectiveness time trends and area and socio economic variations to help inform public health priorities The gold standard for studies of clinical effectiveness is the randomised controlled trial However these trials are carried out on carefully selected subset of patients who usually have more rigorous monitori
58. ge well organised general practices that are highly computerised Therefore they may not be representative of all practices in the locality The introduction of disease registers was one of the first steps towards chronic disease surveillance in primary care by providing information on the prevalence of disease in the population But setting and maintaining a disease register is also very resource intensive For the present however major obstacles remain in trying to use this data obtained from general practice computer systems The principal limitation is that not all GPs are currently recording details of their consultations or other encounters between patients and the NHS on their practice computers Even where GPs are working in paperless practices recording all clinical and administrative information about patients there is no standard method for reporting on the quality of the data recorded Hence the accuracy and completeness of the data may vary widely between practices Problems also remain in trying to extract comparable data from the different computer systems currently on the market Furthermore practices do not uniformly collect socio economic and ethnicity data or risk factor data Practices need to meet specific standards for recording information on morbidity and healthcare use In the longer term the number of practices that can supply high quality information could be increased by better integration of information systems bet
59. gle studies only provide information at one point in time and cannot be used to describe changes in disease prevalence over time The limitations associated with small studies can be overcome by using data from large population based databases of primary care data such as the General Practice Research Database or from information derived from projects such as Primary Care Information Services PRIMIS or the Weekly Returns Service of the Royal College of General Practitioners RCGP see Appendices It is possible to calculate age specific and age standardised disease prevalence rates from these data sources and thus help to monitor national trends in rates of disease prevalence Needs assessment depends on the ability to quantify risk factors diseases and population subgroups This requires the use of routine NHS information systems which general practices may not have access to Most practice computers are set up to record clinical activity and perform only routine administrative tasks Practices that plan to commission health services under the practice based commissioning scheme will require information from public health departments in PCOs on the prevalence of disease effectiveness of treatments and utilisation of health services to allow them to plan health services effectively for their practice population Recent developments in primary care data such as the OOF the Quality Management Analysis System MAS and the Quality Prevalenc
60. h A guide to foundation trusts London Department of Health 2002 14 Marshall MN Shekelle PG Leatherman S et al Public disclosure of performance data learning from the US experience Quality in Health Care 2000 9 53 57 15 Aylin P Best N Bottle A Marshall C Monitoring of mortality rates in primary care London Imperial College of Science Technology amp Medicine 2003 16 Department of Health Tackling health inequalities a programme of action London Department of Health 2003 17 Hunter Du Killoran A Tackling health inequalities turning policy into practice London Health Development Agency 2003 18 Majeed FA Chaturvedi N Reading R Ben Shlomo Y Equity in the NHS Monitoring and promoting equity in primary and secondary care BMJ 1994 308 1426 29 19 Hippisley Cox J Pringle M Inequalities in access to coronary angiography and revascularisation the association of deprivation and location of primary care services Br J Gen Pract 2000 50 449 454 20 McCormick A Fleming D Charlton J Morbidity statistics from general practice Fourth national study 1991 1992 London HMSO 1995 21 Home Office Death certification and investigation in England Wales and Northern Ireland The report of a fundamental review 2003 London HMSO 2003 22 Majeed A Bindman AB Weiner JP Use of risk adjustment in budget setting and performance measurement in primary care how it works BMJ 2001 323 604 607 23 Majeed A Bindman A
61. hat it provides information on consultation rates and disease patterns by socio economic and ethnic group For this reason despite the length of time that has elapsed since the 1991 1992 survey was carried out interest in its findings remains high But its relevance is waning 7 Primary Care Information Services Primary Care Information Services PRIMIS originated from a pilot project called Collection of Health Data from General Practice By April 2000 PRIMIS was more widespread PRIMIS was funded by the NHS Information Authority It is now managed by Connecting for Health The service is led by the Division of General Practice at the University of Nottingham The aim of PRIMIS is to provide education training and analysis to local facilitators to help them assess the data quality of practice systems and to assist practices in using clinical computer systems PRIMIS works with over 200 local information facilitators in England who cascade their knowledge and skills to around 3000 practices Data recorded as part of routine clinical activity are extracted from the computer systems in general practice using MIQUEST queries The queries are designed to examine data quality including completeness accuracy and timeliness The analyses are fed back by facilitators to each Appendices practice to help practices examine and improve their data quality and clinical practice PRIMIS offers a Comparative Analysis Service that provides clinica
62. hers have demonstrated a b the majority of health care both formal and informal takes place in the community More NHS contacts occur and more NHS prescriptions are written in primary care than in any other setting It is no surprise therefore that enormous quantities of data are routinely collected in thousands of separate general practices the length and breadth of the United Kingdom every day These population based data on individuals from birth vary but include information on symptoms investigations diagnoses referrals treatment and outcomes and therefore represent an incredible resource for research education audit quality management service development and planning In an aggregated form this information could provide the most complete picture about the health of the national population as is conceivably possible Furthermore the rapid and widespread uptake of computers in primary care initially for patient registration and repeat prescribing and increasingly to replace all paper records means that these data are now more accessible than ever However much of the evidence we actually use to inform policy continues to be derived from secondary or tertiary care Although there have been some notable advances in knowledge that can be attributed at least in part to data from primary care witness the fetal origins hypothesis arising from the meticulous ledgers of midwife Ethel Burnside and the use of the network of VAMP computers t
63. ional Statistics ONS including tracing flagging migration cancer registration and the longitudinal study although the practical aspects of this remain to be determined SUS will replace everything that exists in the NHS wide clearing system and provide online reporting analysis and extracts A key function of SUS will be to support the Department of Health s policy of payment by results derive dominant Health Related Groups HRGs calculate 0515 2 and provide standard reports Chapter 3 Using data from primary care to Improve health Many organisations and individuals make use of data from primary care The key use of data recorded by primary care professionals is in the clinical care of patients Primary care data have also been used for purposes deemed as indirect care such as the organisation of health services The vast majority of data collected in primary care are by products of administrative activity for example General Medical Services GMS data and more recently the QOF in the new GP contract However there are several specific primary care databases The most widely used is the General Practice Research Database GPRD which has mainly been used for epidemiological and health services research providing information on trends of disease prevalence and prescribing patterns Prescribing Analysis and Cost PACT data provide most of the information on community prescribing a by product of reimbursing pharmaceutical con
64. ire data from primary care to develop sampling frames for clinical trials to provide follow up data on patients in trials and data for other types of research Abbreviations Abbreviations ACG Adjusted Clinical Groups BMI Body Mass Index COPD Chronic Obstructive Pulmonary Disease DH Department of Health DRG Diagnostic Related Group GHS General Household Survey GMS General Medical Services GPRD General Practice Research Database HES Hospital Episode Statistics HRG Health Related Group HSE Health Survey for England ICD International Classification of Diseases LSP Local Service Provider MCA Medicines Control Agency MIQUEST Morbidity Information Query and Export Syntax MSGP4 Fourth Morbidity Survey in General Practice NCRS National Health Service Care Records Service NHAIS National Health Applications and Infrastructure Services NPC National Prescribing Centre National Programme for Information Technology now Connecting for Health NSF National Service Framework ONS Office for National Statistics PACT Prescribing Analysis and Cost PCO Primary Care Organisation PPA Prescription Pricing Authority PRIMIS Primary Care Information Services PSU Prescribing Support Unit QMAS Quality Management and Analysis System OOF Quality and Outcomes Framework OPID Quality Prevalence and Indicator Database RCGP Royal College of General Practitioners SHA Strategic Health Authority SUS Secondary Uses Service THIN The Health I
65. iseases are more prevalent among different ethnic groups due to genetic differences As the population ages more people will have more than one chronic condition This will have an effect on how people use services and how services should be configured to address population need Information on co morbidity is not available from QOF data There are limitations in the recording of risk factors within QOF for example the prevalence of smoking and obesity within practice populations This would provide a proxy figure for the general population and help decision makers anticipate future needs and develop services accordingly There are limitations with QOF data due to the diagnosis and coding of disease and the completeness of practice disease registers There are no standard methods for reporting the quality of data recorded or the completion and accuracy of data or the recording of morbidity Chapter 4 data in primary care This may be overcome by requiring practices to be involved in a data accreditation scheme he development of guidelines for using Read codes for diagnoses may reduce the wide variations in coding among practices computerised for a long time The Department of Health decides the codes that are to be used for defining the disease entries t could be argued that under the new contract GPs have a perverse incentive to use incorrect codes to improve the financial payments that they receive This could be identified through d
66. it Support Programme for Diabetes profect 31 However this would require computer hardware and software systems to be re designed The analyses of primary care data are mainly limited because of issues of data completeness the difficulty in linking data and follow up over time longitudinal study Statistical analysis of data needs to consider issues such as inter and intra practice variation Although the errors associated with these are limited when using larger databases compared with smaller datasets expertise is required in working with large primary care databases Chapter 6 6 1 What is the role of a survey Data provided by surveys can help identify specific problems in the delivery of healthcare services or the health status of individuals Surveys can also be useful for determining patients view about the care that they receive the quality of communications between patients and professionals and in assessing patients physical and psychosocial function as a result of an intervention Surveys are cross sectional and provide information at one moment in time Surveys may be used as part of surveillance However there are concerns with the accuracy of surveys the resources that are required to carry them out and the timescales of data collection Additionally there are methodological issues with how patients are sampled and the significance of people who do not respond to the survey and thus the generalisability of results
67. k of general practices to take part in practice based research The Network Board approves all research studies Initially 72 practices were recruited but some practices dropped out due to problems with committing time to research Practices are paid 1 000 to collaborate in a research project at least once a year All practices had to reach a specified standard in their recording of diagnostic lifestyle prescribing and referrals data The accuracy and completeness of the clinical databases in these studies has enabled cross sectional and case control studies to be undertaken Although practices involved in Trent Focus are representative of other practices in Trent in terms of population and morbidity this may not apply to the rest of England 29 10 Practice based disease registers Before 1999 and the introduction of national service frameworks across clinical areas the first for coronary heart disease the use by practices of disease registers was sporadic The introduction of national service frameworks led to a widespread adoption of disease registers in primary care Practices had to identify the number of patients registered with the practice with a particular condition for example coronary heart disease This was not as simple as it was first made out Patients had to be coded with a diagnosis but many patients were either not diagnosed or inaccurately diagnosed The data for such registers tends to be available most readily from lar
68. land is the main source of routine data on the prevalence of obesity in England Another source is the National Diet and Nutrition Survey The majority of recording of BMI among adults in general practice is ad hoc Obesity levels are increasing among boys and girls Government targets have exposed the lack of information on the extent of childhood obesity Smoking is the single most modifiable risk factor for ill health In primary care the recording of smoking status on general practice computers is variable Smoking data are recorded on separate databases either within hospitals or primary care organisations and are rarely linked t is important to link the varied sources of data on smoking improve the extraction of smoking data from GP systems and to take into account in analysis the recording of smoking status over time Chapter 6 A surveillance system in primary care e There will be differences in the requirements of potential users of primary care data e There is a need to improve the quality of electronic medical records in primary care in particular the recording discipline of doctors e The decisions made as to what data should be collected by a primary care data system should take into account the burden of data collection for individual practices Aanins Aypiquouw Mau e Jo siseq y se pasn aq pino2 u w e 1 UO pnioul jealBojoiqosoiw wa shs
69. lly focussed analysis and feedback on key areas such as heart disease asthma epilepsy and severe mental illness 8 Morbidity Information Query and Export Syntax Morbidity Information Query and Export Syntax MIQUEST was developed between 1992 and 1994 a project jointly funded by the UK NHS Executive Information Group and the former Northern Regional Health Authority The aim of the project was to develop a Windows software system to extract and aggregate data from different GP computer systems Currently MIQUEST is the standard within the NHS for data extraction accreditation MIQUEST has a very strong security mechanism for protecting patient confidentiality An external enquirer that is anyone outside the practice using MIQUEST is unable to extract data with strong patient identifiers from a practice system PRIMIS facilitators are able to extract data authorised by practices following the signing of a confidentiality agreement between practices and their PCO The practice controls the access and release of data extracted using MIQUEST query sets can only be run with a practice s approval and response files can only be transmitted externally with their express approval PRIMIS uses MIQUEST as the preferred approach in supporting analyses of data quality in practices and extracting data for comparative analyses on clinical topics such as coronary heart disease and diabetes mellitus It is also used by Trent Focus and by some other rese
70. many cases the patient record extends from birth and includes details of a patient s diagnoses management and health outcomes e The introduction of a new GP contract with a Quality and Outcomes Framework requires general practices to routinely record detailed information on clinical management In the UK most patients experience of healthcare is within primary care Although over 80 of general practitioners GPs use computers to record patient information information on the consultations between patients and GPs are not routinely collected either nationally or locally Traditionally the main function of information systems in general practice has been to provide information for general practitioners and other members of the clinical team to use in day to day clinical care The data have also been used for patient registration and more recently to help with payments made to practices under the new GP contract The NHS aims to provide cradle to grave care and universal coverage of primary care services by GPs This means that data collected in general practice is population based and in many cases the patient record extends from birth and includes details of a patient s diagnoses management and health outcomes By using primary care data it is possible to extend the understanding of the natural history of illness including access to healthcare by patients The patterns of care in general practice such as rates of prescribing a
71. me visit or a telephone consultation and with whom either a GP or practice nurse GHS is widely acknowledged as useful in examining the rates of consultations with GPs by patient age and in providing a snapshot of annual workload But data from GHS does not accurately monitor trends in GP consultation rates The University of Kent calculates for the Department of Health the cost of a consultation It takes into account the average consultation time and travel time using 1992 data Databases such as OResearch see Appendix 14 may be able to provide more information on the types of consultation than the current GHS once data validity is verified The QResearch database covers 6 of all patients registered with GPs and nearly 500 practices spread throughout the UK The sample data can be re weighted to make it more representative of the UK population 6 Fourth Morbidity Survey in General Practice The Fourth Morbidity Survey in General Practice MSGP4 was a prospective cohort study of around 500 000 patients 1 of the general population registered with 60 volunteer general practices in England and Wales undertaken in 1991 92 Earlier surveys were carried out in 1952 1971 72 and 1981 82 The aim of the survey was to examine the workload and pattern of disease in general practice in relation to patient age sex and socio economic status Many of the practices contributing to MSGP4 continue to provide data for the RCGP s Weekly Returns Servic
72. mprovement Network UKCRC United Kingdom Clinical Research Collaborations VS Vital Statistics erpho J Eastern Region Public Health Observatory Institute of Public Health University Forvie Site Robinson Way Cambridge CB2 2SR t 01223 330353 f 01223 330345 e enquiries rdd phru cam ac uk www erpho org uk January 2006 ISBN 1 904389 09 0 Typeset and printed by Piggott Black Be
73. nd outpatient referrals vary widely but there is little understanding as to why these variations exist There are increasing requirements for GPs to record a wide range of data electronically following the introduction of a new GP contract with a Quality and Outcomes Framework QOF For the first time this requires general practices to record detailed information on clinical management Since these data are linked to payments to practices it is hoped that this will lead to improvements in the accuracy and quality of electronic health records and information in general practice The aim of this report is to e summarise the main systems of data collection in primary care for practitioners e describe the strengths and limitations of existing data sources e examine the potential of new sources of primary care data In this report we focus mainly on the care given by general practice teams and acknowledge the care given by community nurses midwives dentists opticians pharmacists and professionals allied to medicine We use general practice and primary care synonymously We list the main databases and data sources in primary care that are accessible to practitioners see table on pages 6 to 12 with further details in the Appendices We do not distinguish between data collected for direct patient care and that collected for indirect or secondary use We consider the use of primary care data in planning improvements in both individual patien
74. nd secondary care there are large differences in the way in which doctors practice medicine However without taking into account differences in case mix we do not know whether such differences are justified clinically Risk adjustment can help correct such variations for underlying differences in population case mix and thus could lead to performance measures for providers that are fairer and more accurate that the unadjusted measures available now In the USA risk adjustment methods were developed to understand variations in populations A similar development of methods of risk adjustment in the UK using accurate primary care data may contribute to understanding and reducing the unacceptable variations in quality of care Primary care data currently have a limited role in resource allocation in the NHS Prescribing budgets to PCOs are determined using PACT data The current allocation formulas do not use morbidity data from primary care to help allocate NHS resources either to PCOs or to general practices This may be an area of future development Several diagnosis based population risk adjustment models have been developed Diagnostic Related Groups DRGs are a classification initially developed at Yale University DRGs have been used by the Medicare system in the US since 1983 for the reimbursement of health service charges They enabled a prospective payment system to develop with the objective of controlling charges more strictly The mo
75. ng and follow up than takes place in routine clinical practice Therefore they may be atypical of patients and clinical management in the general population But policy makers and clinicians need to know whether the benefits and risks of treatment found in trials are likely to be observed in routine clinical practice Using data from primary care may help in determining whether benefits will translate to the wider population However before researchers can use primary care data there are a number of requirements to improve access to primary care data to improve the range of data that is collected e to link primary care data with other sources of data to develop the recording of socio economic status at a patient level in primary care to improve the quality of data Furthermore new methods need to be developed to rapidly evaluate health policy and clinical effectiveness of interventions using primary care data These observational studies are common in the USA but are carried out less frequently in the UK New research initiatives such as Biobank aim to look more closely at linking datasets to the aetiology of conditions and propose treatments or interventions that would be possible at a system level Biobank will aim to use data from clinical records to record follow up information on patients Chapter 4 Chapter 4 QOF data and QMAS Key messages The national OMAS database currently holds indicator data for 8575 practice
76. nings to support decision making e Identifying patient groups that may benefit from health promotion and preventive medicine e Improved communication and patient follow up especially across health care organisations for example in the tracking of laboratory investigations e Better chronic disease management clinicians by using templates and protocols can ensure that patients with chronic disease can be monitored effectively Practice organisation practices with high quality data can improve cervical cytology recall improve letters of referral ensure claims and payments are made Supporting clinical governance assessment and improvement of the quality of clinical care requires good data Supporting effective commissioning and healthcare planning by PCOs and more recently practices that require operational clinical data on areas such as workload quality of care and health care burden Providing data to the wider NHS a great deal of the work of general practice was hidden because either the data were unavailable or there were no requirements for data in contrast to the hospital sector The use of health data especially primary care data can be classified into a number of broad themes e Needs assessment health service planning and commissioning e Regulation accountability and performance management e Clinical governance and quality improvement e Monitoring health inequalities e Monitoring healthcare use e Monitoring death rates
77. ns Note 7 Introduction to 2004 2005 QOF data will be published by the Eastern Region Public Health Observatory in Spring 2006 Chapter 5 Case studies eoo 5 1 Primary prevention and public health surveillance Case A How do we monitor the prevalence of adult and childhood obesity Key messages e Obesity is predicted to have an effect on the health of the population equivalent to tobacco smoking The Health Survey for England is the main source of routine data on the prevalence of obesity in England Another source is the National Diet and Nutrition Survey The majority of recording of BMI among adults in general practice is ad hoc Obesity levels are increasing among boys and girls Government targets have exposed the lack of information on the extent of childhood obesity Obesity is predicted to have an effect on the health of the population equivalent to tobacco smoking In the UK the numbers of adults and children who are overweight and obese are rising markedly In 2002 22 of men and 23 of women were defined as obese with older people and the less wealthy more likely to be obese 24 Obesity is usually defined in terms of the Body Mass Index BMI and a BMI of over 30kg m2 is classed as obese However a more refined measure of obesity is a high waist to hip ratio which describes the distribution of fat in the body and is shown to be associated with high r
78. nths DM 10 The percentage of patients with diabetes with a record of neuropathy testing in the previous 15 months DM 11 The percentage of patients with diabetes who have a record of the blood pressure in the past 15 months DM 12 The percentage of patients with diabetes in whom the last blood pressure is 145 85 or less DM 13 The percentage of patients with diabetes who have a record of micro albuminuria testing in the previous 15 months exception reporting for patients with proteinuria DM 14 The percentage of patients with diabetes who have a record of serum creatinine testing in the previous 15 months DM 15 The percentage of patients with diabetes with a diagnosis of proteinuria or micro albuminuria who are treated with Angiotensin Converting Enzyme inhibitors or Alpha 2 antagonists DM 16 The percentage of patients with diabetes who have a record of total cholesterol in the previous 15 months DM 17 The percentage of patients with diabetes whose last measured total cholesterol within the previous 15 months is 5mmols or less DM 18 The percentage of patients with diabetes who have had influenza immunisation in the preceding 1 September to 31 March The indicators predominantly focus on measures of risk factors such as smoking which are modifiable by changes to lifestyle or in recording BMI levels thus identifying patients that are overweight Other indicators such as cholesterol measurements may be partly modifiable by lifestyle changes
79. o Identify serious adverse drug reactionsd this resource has been underused and perhaps occasionally even misused Data collected in primary care has strengths and limitations and it is important for those collecting analysing and interpreting it to be aware of these Algorithms have been developed to clean and code computer data yet every practitioner whose surgery list has run thirty minutes late will know that the quality of data entry can vary While the enormous size of pooled practice datasets can overcome error when assessing the strength of associations between different variables it is not so easy to deal with bias Clearly primary care data is of enormous value in helping practitioners to do simple things well for many people few of whom actually feel ill which remains the hallmark of chronic disease care However quantifying the population frequency of important health behaviours such as physical activity or the association between such behaviours and disease endpoints will continue to require specialised studies There is a danger that the availability of a wonderful resource such as a large network of practices downloading medical records daily can start to drive the formation of research questions which is undesirable in the longer term In developing the academic discipline of general practice there has been an understandable rush to the randomised trial The availability of data from large cohorts of patients in primary care should
80. of Health PCOs report on a quarterly basis the number of people seen by the stop smoking service All hospitals have to collect information on the proportion of pregnant women who smoke Smoking data are also recorded as part of Child Health Surveillance systems The smoking status of parents for every newborn is recorded on the discharge sheet But often this information is frequently not used at a local level Unfortunately these data are recorded on separate databases within hospitals and PCOs In primary care the recording of smoking status on general practice computer systems is variable The recording of smoking status tends to be high for new patients as this question is part of the registration health check The recording of smoking status among patients with chronic conditions is also part of the new GP contract quality and outcomes framework Data collected within primary care and the community setting may help evaluate interventions that will be effective in reducing smoking prevalence The smoking prevalence among children of school age is not known Usually this information is collected by surveys along with information on drug and alcohol use among school children t is proposed that GPs record the percentage of people aged 15 to 75 years who smoke These data may then be analysed by five year age sex groups and compared with existing population and survey data Furthermore details on the recording of the provision of stop smoking
81. of more clinically focussed performance indicators which can be aggregated at a regional and national level The Department of Health and the Healthcare Commission which exists to promote improvement in the quality of healthcare in England are responsible for developing a new performance framework for PCOs They are all likely to use data from primary care especially QOF data in the task of performance monitoring NHS and non NHS organisations The results from these new performance measures will be vital in assigning star ratings and Foundation Status to organisations 13 NHS organisations deemed to have under performed are likely to face sanctions Therefore it is important that performance indicators provide an accurate and valid assessment of quality of care across key areas and that information is interpreted appropriately as perverse incentives may operate National sources of primary care data such as GPRD and PRIMIS and data from GMS QOF can provide information on how practices vary The benchmarking of PCOs and general practices through the use of national and local standards may help identify areas of clinical practice requiring improvement The advantage of using performance indicators is in their potential effect of altering behaviour in a beneficial way In New York State the publication of indicators of death rates among patients who underwent cardiac surgery was initially followed by a fall in patient death rates 14 n England r
82. ot available or useable across other parts Although there are now improvements in the recording of clinical information by GPs there are still many gaps for example in the recording of results of diagnostic investigations This is partly due to the opportunity cost and resource implications for practices in recording clinical data on general practice computer systems for example in lengthening the consultation time between doctors and patients In many cases the audits have used tools and techniques developed by PRIMIS Some practices are developing the use of clinical data together with guidelines and decision support software Such links are likely to become more widespread and may help improve the process of clinical governance 3 4 Monitoring health inequalities Reducing inequalities in health status among the UK population is a key priority for the NHS 16 17 The NHS has two roles in improving the delivery of healthcare services ensuring equity of access and in improving population health through prevention Currently there are gaps in understanding the extent to which primary care use is related to need among differing socio economic and ethnic groups or whether there exists age sex or ethnic discrimination Research shows there is an association between deprivation and access to primary healthcare the inverse healthcare law which states that patients with the greatest need for healthcare are least likely to access care Many small studie
83. ows for differences between practice prevalence observed and population prevalence expected to be examined Along with the collection of information on the organisational aspects of the QOF framework this may help us understand the factors that are necessary to provide high quality general practice QOF data signal new possibilities to investigate variations in primary care and test hypotheses e g do practices with larger list sizes achieve higher OOF scores and thus better care To understand the potential of QMAS data to provide information on the prevalence of disease at a local level it is important to understand the accuracy of the practice list size and the population characteristics of those registered with the practice Inaccuracies in the list size will alter the denominator for calculating the prevalence and thus this may lead to underestimates or overestimates of disease prevalence Exception reporting is also important to understanding MAS data Exception reporting was introduced to prevent practices being penalised under the quality and outcomes framework for factors for which they had no control The exception code is used so as to not affect the practice s quality point score There are a number of reasons for exception coding and therefore the clinical indicators need careful interpretation A denominator for a specific clinical indicator may not be the same as the disease register size to which the clinical indicator relates
84. pleteness and accuracy of coding Disease registers are most useful in situations where disease or risk factor status does not tend to change over time the diagnosis of disease needs to be consistent and based on a robust diagnostic test Furthermore a register is of use when there is a requirement for ongoing health care for example retinal screening among patients with diabetes Registers can be used for many purposes to monitor temporal trends follow up patients help compare treatment outcomes help undertake evaluation of services and study the causes of diseases and help organise services for patients If case ascertainment is high prevalence and incidence rates can be calculated and analyses of risks and aetiology can be explored using individual and area characteristics With follow up data outcomes such as survival rates for cancer can be measured If registries are maintained over time they can produce evidence of change as in epidemics or in the effectiveness of interventions and be of use in surveillance Registers can be used to assist in the management of chronic disease management in clinical settings help trigger the follow up care for people with for example diabetes or asthma within general practice Registers can form the basis for clinical audit and quality improvement and for providing services such as child protection registers There are several limitations of primary care disease registers hey will not tell you
85. population Because many chronic illnesses do not result in death mortality rates are not always a good measure of the health of a population Similarly a generic measure of chronic illness based on self reports may also be an unreliable measure of health status By contrast risk adjustment models based on all conditions treated by primary and secondary care physicians take into account the full spectrum of illnesses in a population Hence they incorporate chronic illnesses such as arthritis or epilepsy that lead to considerable population morbidity but that are often not recorded on death certificates PCOs and primary health care teams need local data to help them plan local health services and monitor the quality of care they provide They can make use of national data for example to provide baseline estimates of disease prevalence or to compare local treatment patterns against national patterns However their principal requirement is always likely to be for local rather than national data 3 10 Financial flows and payment by results Primary care data are likely to be used to develop accurate pricing of healthcare interventions by different providers within primary and hospital care As part of the Department of Health s policy of introducing payment by results within the NHS national tariffs for primary care interventions and procedures are likely to be set The aim of the new financial system is to provide a transparent rules based system
86. prospective studies and the results disseminated and action taken to protect the public health This can take place long before any findings would be available from prospective studies because they use data that are already collected Large primary care based databases can provide information on the long term follow up of new and established drugs and help to answer questions about the safety of drugs Primary care databases such as GPRD have mainly been used for pharmaco epidemiological studies This research has been driven by the pharmaceutical industry which has a major interest in this area They have funded many of these studies and determined the research agenda It is important that public health priorities should also drive this type of pharmaco epidemiological research 3 9 Resource allocation risk adjustment and case mix Governments of all health care systems are looking to provide services that are cost effective of high quality and are responsive to people In the UK the quality of primary care has varied enormously and much of this variation has been largely unexplained Health care systems around the world are all trying to ensure that resources are used appropriately and as part of this process they are trying to reduce unacceptable variations between providers For example the NHS Plan stated that it promises better performance and accountability systems to reduce variations in services across England In both primary a
87. r between five and ten years In the QOF there are 18 indicators for diabetes care shown below QOF clinical indicators relating to diabetes mellitus DM 1 The practice can produce a register of all patients with diabetes mellitus DM2 The percentage of patients with diabetes whose notes record Body Mass Index BMI in the previous 15 months DM3 The percentage of patients with diabetes for whom there is a record of smoking status in the previous 15 months except those who have never smoked where smoking status need be recorded only once since diagnosis DM 4 The percentage of patients with diabetes who smoke and whose notes contain a record that smoking cessation advice or referral to a specialist service where available has been offered in the last 15 months DM 5 The percentage of diabetic patients who have a record of glycosylated haemoglobin HbA1c or equivalent in the previous 15 months DM6 The percentage of patients with diabetes in whom the last HbA1c is 7 4 or less or equivalent test reference range depending on local laboratory in last 15 months DM7 The percentage of patients with diabetes in whom the last HbA1c is 10 or less or equivalent test reference range depending on local laboratory in last 15 months DM8 The percentage of patients with diabetes who have a record of retinal screening in the previous 15 months DM9 The percentage of patients with a record of the presence or absence of peripheral pulses in the previous 15 mo
88. re services and for organising referrals for specialist care Consequently the medical records held by general practitioners include details of all diagnoses and prescribed drugs in many cases extending from birth In an increasing number of general practices these records are held in electronic format and are potentially available for extraction and analysis Electronic data from general practice therefore offer unique opportunities to plan and monitor health services measure the quality of care provided by the National Health Service and undertake population based research The strengths of the data collected from general practice are that they are population based and not derived from an unrepresentative subset of the population that most contacts with the NHS take place in primary care and they contain information on illness treatments outcomes and use of healthcare services Now that the NHS is investing heavily in information technology through Connecting for Health formerly the National Programme for Information Technology in the NHS careful thought needs to be given to how to improve the quality of electronic health records in general practice to derive maximum benefit from the data they contain For example there are now opportunities to link information from primary care records with information from other sources such as hospital episode statistics to provide a more comprehensive record of the clinical care experienced by patients
89. rganisations SUS will provide facilities for quality reporting standard analyses user analysis and extraction 16 MEMO MEMO was originally set up in Tayside in Scotland to undertake pharmaco vigilance studies using datasets on dispensed prescribing hospitalisation and death certification MEMO is able to record link a wide range of primary care secondary care and pharmacy datasets This information although collected in Scotland may be applicable to improving health services and public health in England as the demographics of the Tayside population are representative of the UK and Europe 17 The Health Improvement Network EPIC and n Practice Systems two commercial organisations have developed The Health Improvement Network THIN primary care database to facilitate the use of NHS electronic databases of primary care records for medical and pharmaceutical research It is a new database of anonymised clinical records EPIC has a long history of using electronic medical records in research and In Practice Systems in developing and supplying the Vision general practice computer system Data collection started in January 2003 Over a hundred practices using Vision software have joined THIN Most practices have recorded several years of data on their system and where contributing practices had previously used VAMP systems data entries extend to 1987 An independent research panel will review all research proposals and THIN data will
90. ry Uses Service SUS is part of the NHS Care Records Service NCRS The service aims to provide pseudo anonymised patient data for purposes defined as other than direct clinical care This includes planning commissioning public health clinical audit benchmarking performance improvement and clinical governance The aims of SUS are to reduce the burden on the NHS of collecting abstracting and submitting data Furthermore to provide national comparators increase timeliness and quality of data available to the NHS and its partners and access to national transparent data from the NHS and other organisations e g ONS As a result of better more timely public health information the quality of care for patients can be improved Appendices Initially SUS will take on the data flowing through the current NHS Wide Clearing Service NWCS and support the implementation of Payment by Results Over time other data sources will be included including cancer waiting times clinical audit information and central returns It will also include data from non patient based sources e g ONS workforce data and collect data from the NCRS patient records right across the care pathway The service will also generate some central returns and on line access to analytical tools and services will be also be available for research The information within SUS will have rigorous access controls providing online access to the NHS and to other related agencies and health o
91. s have reported difficulties in access to primary care for patient groups such as the homeless asylum seekers refugees and minority ethnic groups 8 Data collected in primary care because of the universal coverage of primary care services can provide important information on the morbidity of the population and how this varies among different socio economic and ethnic groups Collecting this information will allow the NHS to systematically monitor whether there is equitable access to care Previously the only main sources of patient level socio economic data in primary care have been the national morbidity surveys in general practice the last carried out in 1991 1992 Accurate and systematic allocation of socio economic information to individual patient records has been undertaken as part of these surveys Such data are not available in routinely collected medical records Subsequent studies by the ONS and the Department of Primary Care University of Nottingham have used practice postcodes as a proxy measure of deprivation to examine the patterns of healthcare use and health outcome 19 Analyses using this approach usually show that the prevalence of chronic disease is highest in the most deprived areas and is associated with the lowest rates of treatment A recent study showed the prevalence of coronary heart disease was highest in deprived areas and lowest in affluent areas whilst the use of lipid lowering therapy showed the opposite pattern Furth
92. s in England QMAS is a national web based software tool developed for implementing the new GP contract Data from practices are aggregated to maintain patient confidentiality and for every practice a set of quality QOF scores is calculated Information is collected on the number of patients with a particular chronic disease condition and on the care they receive There is potential to link QOF data with other data collected in primary care and with other data sources This would allow practices to be compared adjusting for differences in the underlying population to help understand differences in the quality of care provided QOF data have limitations and these include not being able to calculate age standardised prevalence rates adjust for socio economic and ethnicity differences in the population or adjust for inconsistencies in diagnosing and coding of disease conditions between healthcare professionals The Quality Management and Analysis System QMAS is a national web based software tool developed in response to the introduction of a new GMS contract in April 2004 by Connecting for Health lts aim is to extract data from GP practice systems The Quality Prevalence and Indicator Database OPID held by the Health and Social care Information Centre in Leeds aims to improve access to QMAS data by users in the Department of Health and NHS The Information Centre will undertake national analyses for example of disease prevalence
93. scribed drug costs They cannot be linked to demographic or clinical data on patients They cannot be used to calculate age and sex specific prescribing rates They cannot be used to look at prescribing rates for specific conditions PACT does not include private prescriptions or prescriptions that the patient does not have dispensed The number of items prescribed is not always an accurate measure of the amount of a drug actually prescribed t may be possible to expand the use of PACT data by including diagnostic data on the prescription issued 13 Ouality Prevalence and Indicator Database The Quality Prevalence and Indicator Database OPID is being developed by the new Health and Social Care Information Centre in Leeds The aim of P D is to widen the access to QMAS data by users in the Department of Health and NHS OPID and the Information Centre aims to provide analytical support and undertake national analysis on prevalence using QMAS data and other data sources They are to publish data in readiness of the freedom of information legislation To access the OPID data users will have to inform the OPID team of the proposed use of data or information from OPID which is then considered for approval by the OPID Gateway Committee Approved requests lead to detailed user specifications The OPID team and the user will sign an agreement on the provision and use of data 14 OResearch The aim of OResearch is to develop and
94. se By monitoring the effect of preventative measures such as exercise diet smoking prevention and alcohol it may be possible to develop more effective interventions PCOs are likely to examine cost data to determine current levels of expenditure on diabetic care Potentially larger practices may have scope to make savings and primary care data especially QOF may contribute to determining the cost effective of primary care based diabetic services as opposed to services delivered by hospitals or other providers Chapter 6 Chapter 6 A surveillance system In Key messages e There will be differences in the requirements of potential users of primary care data e There is a need to improve the quality of electronic medical records in primary care in particular the recording discipline of doctors e The decisions made as to what data should be collected by a primary care data system should take into account the burden of data collection for individual practices Publications such as Key Health Statistics from General Practice used primary care data to report national estimates of chronic disease prevalence and management and thus helped to raise the importance of data collected in primary care 2 More recently it has been the introduction of a new GP contract and the associated requirements of practices to demonstrate their performance that has raised further interest in using primary care data to help plan monitor and improve the
95. st widely used primary care based risk adjustment method is the Adjusted Clinical Group ACG system developed at John Hopkins University in Baltimore USA 2223 The ACG system clusters diagnoses into clinically meaningful categories and for individual patients gives a composite measure of health status to help predict the patient s future use of health services Its objective is to help ensure that the morbidity of individual patients is accounted for in allocating budgets A patient with heart disease and diabetes would be placed in a higher category than a patient with ischaemic heart disease In the USA risk adjustment is now starting to be used to adjust capitation or other types of payments to health providers For large populations adjustments for age and sex may be adequate for this purpose but this is not the case for smaller populations such as those managed by small healthcare providers Risk adjustment helps ensure that providers of healthcare who manage patients with more complex medical problems have their budgets adjusted to take this into account The use of risk adjustment systems has also given doctors and providers a powerful incentive to provide more accurate and complete diagnostic data Another use of risk adjustment is to measure the health status of a population The traditional way of doing this in most countries has been to use mortality rates or self reported measures of chronic illness derived from censuses or surveys of the
96. support decision making e Identifying patient groups that may benefit from health promotion and preventive medicine e Improved communication and patient follow up especially across health care organisations for example in the tracking of laboratory investigations A user s guide to data collected in primary care in England Better chronic disease management clinicians by using templates and protocols can ensure that patients with chronic disease can be monitored effectively Practice organisation practices with high quality data can improve cervical cytology recall improve letters of referral ensure claims and payments are made Supporting clinical governance assessment and improvement of the quality of clinical care requires good data Supporting effective commissioning and healthcare planning by PCOs and more recently practices that require operational clinical data on areas such as workload quality of care and health care burden Providing data to the wider NHS a great deal of the work of general practice was hidden because either the data were unavailable or there were no requirements for data in contrast to the hospital sector The use of health data especially primary care data can be classified into a number of broad themes which are discussed in more detail in this section Needs assessment health service planning and commissioning Regulation accountability and performance management Clinical governance and quality
97. t care and population based care The new Health Information and Intelligence Task Force has the responsibility to develop a health information and intelligence strategy as recommended by the Department of Health s public health White Paper Choosing Health We acknowledge that changes in NHS policy continue to take place but we hope this report will be of use to public health practitioners who wish to expand their knowledge and use of data collected in primary care Chapter 2 Connecting for Health Key messages e Connecting for Health aims to centralise the electronic records of 50 million patients e The Health and Social Care Information Centre aims to combine information systems for health and social care In 1998 the Department of Health s strategy Information for Health 3 committed the NHS to setting up lifelong electronic health records for all In 2002 the Government set up the National Programme for Information Technology now called Connecting for Health to overhaul the information technology infrastructure of the NHS by 2010 allocating 6 3 billion to the project The aim was to digitise the 50 million patient records creating a central record of patients care This would allow information to be shared safely across the NHS and for individuals to have access to their own electronic records t would hold a summary of a patient s contact with all care providers and would record the patient s consent for profession
98. ten down by healthcare providers in the medical record and are the data used to diagnose patients and determine treatment plans Drugs bill a term for the allocation of a budget for prescribing Indicator a tool for measuring how well a standard is achieved Local Service Providers those responsible for delivering IT services and support on a local level for five regional clusters of strategic health authorities Patient identifiers aspects of information that can be traced back to the patient typically date of birth name and address Performance the extent to which a defined standard is achieved It is a subjective judgement of a level of performance that should be achieved and may be set nationally or locally Primary care in this report defined as care provided by general practice teams including health visitors district nurses and mental health nurses and excludes care given by dentists pharmacists opticians and midwives n this report general practice and primary care are used synonymously Standard a statement of an agreed level of care Surveillance a system that involves the scrutiny of all aspects of disease occurrence both communicable and non communicable including the monitoring of interventions that may reduce disease prevalence Executive summary Executive summary eoooooooooo o General practitioners are responsible both for providing primary ca
99. than referring them to hospital for treatment Information is collected as to the number of patients within a practice that have a particular chronic disease condition such as coronary heart disease or diabetes mellitus The key information collected by MAS is on the disease register size and the numerator and denominator for each clinical indicator based on a set of defined Read codes No demographic information is collected by MAS The clinical indicators that relate to each chronic condition are measures of process rather than outcome for patients with chronic conditions Some of these process indicators but not all are linked to better health outcome good blood pressure and blood glucose monitoring among people with diabetes mellitus is associated with lower mortality rates Exception codes Practices are allowed to exclude patients from the denominator for an individual clinical indicator through exception coding There are two types of Read exception codes A high level code that excepts from the whole clinical domain for example if a patient has diabetes mellitus and does not wish to attend for review There are also indvidual exception codes for indicators within the clinical domain for example those declining to have a flu injection But when data are entered for individual clinical indicators this overrides the high level exception code for that clinical domain The exception codes cover the following areas e Patients who refuse
100. therapeutic areas such as respiratory drugs Analysed information is fed back at practice PCO SHA and national level But users do not have access to local data As prescribing is heavily influenced by practice demography data are adjusted for age and sex to compare the rates and costs of prescribing in different practices or health authorities GPs generally receive a standard PACT report that contains information on the practices rates and costs of prescribing along with comparative information GPs can ask for more detailed information but this report is often very unwieldy PCOs receive PACT reports on GPs prescribing and can access computerised PACT data through PACTLINE epact net PCO pharmaceutical advisers can obtain information on the prescribing of specific drugs online directly from the PPA PACT data have been used mainly for two purposes First as a financial tool to help health authorities set and monitor prescribing budgets in general practice Second in allocation formulas for prescribing budgets by the Department of Health Before the new GMS contract and its allocation formula health authorities had to use their judgement in allocating the prescribing budget for practices This may in part have led to the development of practice formularies PACT data have also been used in audit and research But there are limitations to PACT data e They provide information mainly on what drugs are prescribed and how much the pre
101. tractors PACT data are also used for analysing prescribing in therapeutic areas identifying expensive drugs developing prescribing indicators setting and monitoring budgets and health services research Researchers have used PACT data to investigate variations and trends in prescribing costs for example between fundholding and non fundholding practices Primary care organisations PCOs can identify practices that prescribe new and high cost drugs compared with older drugs of similar efficacy or drugs that are known to be ineffective using information from the Prescription Pricing Authority However there are key limitations with data collected by GPs e The most heavily used primary care databases have data collected from volunteer practices Volunteer practices are often large practices that are deemed to provide an above average quality of care and therefore are often not representative of all practices e The quality and completeness of electronic health records is highly variable and e nformation on social circumstances and ethnic status is often absent There are several expected benefits from general practices improving information management and data quality These include e Improving patient care within the consultation high quality data about patients at the point of care ensures that there is legibility of medical notes data are complete and comprehensive there are alerts to clinical errors and appropriate warnings to
102. ue pue 03 S pue puejas WayLON S EAA pue Bu3 1nouBnoiu s n bid 00q D TD Q H S8311381d je p w usi Buj3 004G uu01J e ep SuIe u09 edIpaw ul asn 10 ejep aaijoeid jesauabh 10 aseqeyep jojid e e sp 7 Ul 811u39 UO JEUOJU 9482 15190 pue yyjeay Aq padojanaq Gld0 seqe e 10 e3 pul pue aouajeaaig Ayjeng 1281402 45 J0 149 sawon pue Nemo au 10 p312811x8 ejeq sul 81s s 391 7321d 40 wos 19841X8 0 1 80 Bu l99uu02 Aq 100 paseq SVIND was g s s jeuy pue uawa feue Aileno asoding pue inos Ju uido A p e 91 s s yuawdojanap Z xipuaddy 400 Aeio ds saseqejep 18y 0 YUM e ep jo suosuedwog e 91 aas GI s s yju uidoj n p 10 se iV 3A1 e U S 1d 1 aq 100 Aew 31049184 pue SJ99 UNJOA S A 0AU e 6002 41 pales aseqejep Mau AHE H eyep YPIJEA e se SNONIQUIY UOI EUO UI BulpiAoid 10 su0 1e3 duii snowJous BI1u910d e S SS U B NA ge Jo dn A0 04 do gA p Aew q suonlpuoo pue aseasip y pue JUaWUOJIAUa sajAysayl Sauab s juediaiued y US Ml Q Ul y p
103. ue siepun 01 8 qQ8 aq S UO P S H Asoysiy IU E pue jeaipaw pue juaWAo dwia A1s Ji sig 1un o0A UO UO JEUOJU pIA01d VAN SWA SAS dINVA Sn0 A81d y 01 lnqilu05 0 p sn pue ejep Jo sie ALY S8311981d JSON 60101 2 pasiwAuoUY doin3 pue xn ul suoneindod 0 ajqesijejauay sjasejep Aoguuieud pue b Mepuosas Aewud juj 0 lqy e ep jeloos pue yjjeay 3uiqui09 e ep ua1 ed pasiwAuoue opn sd piA04d IAA SHN 324 UO U01198 03 aonpal 0 wiy syjiuans sleak 69 0y U Ml Q pabe sJ l1un oA 000 00G e 89910640 001 0 5 pis e u do d 0007 007 110 8 519 01 uonpindod JO ysabie p inq 0 1391014 yuegorg yn SuI81S S 8211381d U pue 2144 Suol esiueBi0 61 2 om Aq padojanag e NIHL Oman quawanoiduy yyyeay L U011891J11195 ule p pue uoljesijeqidsoy Buiqiu9s id pasuadsip sjaseyep eUIBIO 391U e salpnys oue IBIA o3euueud Jo Busa sisayjodAy 0 ul dn 195 AjjeulBuiQ OINIIN 8183 8010119 JOSIP ULY 19470 s sod nd 104 pasn aq Ol e 0 2 SD 029H 9160 SHN 84 J0 rM s s sn uppuos s asoding pue inos Chapter 1 Key messages e Data in general practice are population based and in
104. ve verified that GPRD data are of high quality are supplied by volunteer general practices and are unlikely to be representative of all UK practices The recording of data is likely to be far more variable A study of 47 general practices all in one PCO in south west London showed a wide variation in the coding of patients with ischaemic heart disease 28 No practice had all cases coded although some achieved coding levels of over 90 and in some practices no cases were coded One problem that GPs face when improving the quality of their computerised data is the lack of standards in recording morbidity data Even among practices that have been computerised for a long time there are wide variations in the coding of clinical information by different practitioners For example there are many codes that are used for heart disease Some researchers have proposed a data accreditation scheme with standards for completion and accuracy of data and guidelines for using codes for diagnoses Standardisation would allow for comparisons of general practices and PCOs and allow data to be aggregated to produce national estimates For example tools such as Morbidity Information Query and Export Syntax MIQUEST a computer program designed to extract information from GPs computer systems are used to help obtain practice data in a standardised way MIQUEST has been adopted as a standard within the NHS for data extraction and practices are now required to be MI
105. ver there are concerns about the accuracy and sensitivity of information with the NHS Care Records Service There are three main steps in ensuring patient confidentiality and appropriate disclosure of data First there has to be a legitimate relationship between the patient and the care professional accessing the patient s record Second professionals should only have access to information as they need it Finally patients may choose to have information placed in sealed envelopes if they do not wish it to be revealed during routine care The rules governing information held in the service are set out in the NHS Care Record Guarantee Patients have a right of access to health information about themselves under the Freedom of Information Act while the Data Protection Act 1998 governs access to the health records of living patients 11 The Freedom of Information Act has implications for health service organisations in how information is made available to the public including patients and carers consumer organisations professionals and professional bodies and local and national government There are other components of Connecting for Health which are described in more detail within the Appendices such as the Secondary Uses Service SUS defined as the use of NHS data for purposes other than direct patient care SUS will include all NHS activity and other non patient record based data It will replace some of the functions of the Office for Nat
106. ween primary and secondary care reducing the need to record data more than once clear guidelines on recording data regular feedback on data quality and incentives for practices to become fully computerised 11 Practice based health promotion data The development of NSFs and the new GMS contract has placed a greater emphasis on recording data on risk factors of disease such as smoking and obesity Choosing Health documented an epidemic rise in obesity levels and the implications for public health Determining the extent of Appendices particular lifestyles or risk factors within a population is an important aspect of assessing population health need As with morbidity data the sources of data on lifestyle are very disparate in primary care The main source of data is from national surveys which are then applied to local populations 12 Prescribing Analysis and Cost PACT data Prescribing Analysis and Cost PACT data are a by product of reimbursing pharmaceutical contractors Information on all dispensed NHS prescriptions is collected by the Prescribing Pricing Authority PPA The information collected includes the name and cost of the drug and the number of items dispensed The drugs dispensed are then used to calculate the cost of each item Drugs are categorised by the section of the British National Formula that they fall in Information is available for individual drugs for categories of drugs such as bronchodilators and for
Download Pdf Manuals
Related Search
Related Contents
Precaution when Shifting to ROM-Read Modes Heizkissen bosotherm 1300 bosotherm 1400 Pioneer F60 Operation Manual 取扱説明書 FreeVia 300 Soleus Air HM5-15A-32 User's Manual User`s Manual OWNER`S MANUAL IKEMI & GENKI QUARK - Linfo Copyright © All rights reserved.
Failed to retrieve file