Briefing June 2013 Issue 266

Hospitals Forum briefing June 2013 Issue 266

The non-executive directors’ guide to hospital data Part four: How to make good use of data – quality and safety, including mortality, activity data, contracting and finance

Key points Understanding your organisation’s data is an essential part of providing effective oversight. But data may not always give you the complete picture • Many indicators can be used and it is important to first understand what data is available, how it is to monitor quality and safety. recorded and what these records are used for. They should not be considered in isolation. This Briefing will help non-executive directors (NEDs) better understand • The quality of indicators is NHS data and how it can be used to determine what is going on in their inextricably linked to the quality hospital. For the purposes of this Briefing we examine data in the acute of data. care setting only. Data is of course collected in primary care by GPs, • The accurate coding of clinical pharmacists, dentists and opticians, but the various datasets are not activity is fundamental, linked by the NHS. particularly when constructing mortality rates. This Briefing looks at how to make good use of data across: quality and safety, including mortality; activity; contracting; and finance. It also • Good clinical coding is also crucial includes a short technical guide on ICD-10, OPCS-4 and healthcare in determining the right level of resource groups (HRGs). payments for healthcare services, and thus defining NHS tariffs. How to make good use non-executive directors need to of data be aware that they have to ask Data, by its nature, is something the right questions in order to that should be handled carefully. establish what the data is telling NHS data is no exception and them. The questions are simple

Produced in association with briefing 266 The non-executive directors’ guide to hospital data: part four

‘There is no single ‘dashboard’ Key questions for NEDs to ask that should be used; the recommendation is that • Do we have a collaborative approach between clinicians and coders to trusts monitor a wide spread ensure the data we hold is accurate? of indicators and consistently • Is there a robust clinical governance process in place regarding the track these over time’ review of mortality to provide assurances to the board?

• Do we have mortality sub-groups in place within our organisation? and straightforward. For instance, do our emergency admissions • Are readmission cases a result of the quality of care we provide or a lack figures include mothers and of community services for the treatment of certain conditions? babies? Without answers to these questions the data cannot be • Are we monitoring the same range of quality and safety indicators interpreted correctly. over time?

We have already highlighted the • Do we benchmark our performance against other providers? sort of questions that can be asked of NHS data in parts one, trusts to make sure they are characteristics which would two and three of this series. We comparing like with like, and peer affect the indicators (such as now turn to how the data can be used to help trusts improve the group comparison shines a light age, diagnosis, co-morbidities quality and safety of the services on the areas where improvements and procedure). Adjustment they provide. can be made. can be made for a number of indicators, including length Quality and safety However, non-executive directors of stay, readmissions and need to be aware that in order mortality. Although it sounds There are many indicators to benchmark data a number of straightforward, performing that can be used to monitor things have to happen. First is clinically credible risk the quality and safety of care adjustment of the raw numbers. adjustment is difficult and there provided by a hospital trust. At its simplest level, this is for the are different ways of adjusting However, there is no single size of the hospital trust and the for risk. The following example ‘dashboard’ that should be used; number of patients that it delivers explains the basics. the recommendation is that care to in comparison to other trusts monitor a wide spread of trusts (turning “raw” scores into Risk adjusted indicators and consistently track rates, for example, the number of readmissions within these over time. events per 1,000 patients). 30 days Although this intelligence is The next stage of refinement is This indicator is the relative risk useful, it is one dimensional. the construction of risk-adjusted of readmission within 30 days. The board can only see how the indicators. Risk adjustment is It is the ratio of the observed trust is performing in isolation. the process of adjusting for risk number of readmissions Benchmarking is required to factors (which might explain to the expected number of assess performance within, and variation in outcome) so that readmissions (taking into across, the healthcare system. comparison can be made. account the various risk factors). There is, understandably, a In essence, the adjustment This ratio is then multiplied growing market for benchmarking is not only for the volume of by 100. The factors that will services. Providers work with patients seen but also key influence the ratio are:

02 briefing 266 The non-executive directors’ guide to hospital data: part four

1. The accuracy of the observed When looking at quality of care, number of readmissions indicators like readmissions ‘Non-executives should be aware This sounds like it should be a within 30 days and length of that a longer than expected simple count: how many people stay might be considered useful. length of stay can be a result of have come back as an emergency It is possible that a higher than poor coding of co-morbidities’ admission within 30 days of expected level of readmissions discharge? However, the rules for within 30 days in a given specialty this indicator, used for the current (such as orthopaedics) indicates • HSMR – Hospital Standardised Mortality Ratio (Dr Foster financial penalty, include people that operations are not being Intelligence). who end up being admitted to carried out as successfully as they a different hospital within the are elsewhere. Likewise, a longer SHMI is the only measure now 30 day period, which your staff than expected length of stay on published on NHS Choices. will not be able to ‘see’ until the a specific ward might indicate central data lets them know. problems with the quality of care All three measures are usually provided – perhaps a few patients expressed as a value of 100. An 2. The way the expected number developed pressure ulcers. index above 100 indicates more of readmissions is calculated deaths than expected, whilst a There is no one single ‘correct’ However, non-executives should be lower index indicates fewer deaths way to predict the expected aware that a longer than expected than expected. The Health & number. It relies on having length of stay can also be a result Social Care Information Centre a good understanding of the of poor coding of co-morbidities. publishes SHMI slightly differently. way services are provided. For This means hospital data will It uses 1.00 as the average, which instance, are there any sub- not record how sick the patient is often multiplied by 100 to allow groups who should be excluded is. Alternatively, it could signal a comparison. for specific reasons? Looking pathway issue where a consultant at readmissions, many cancer might be making a judgement The NHS Medical Director has treatments are not carried out to a based on years of practice such made it clear that mortality fixed timetable so patients might as: “I always keep my patients in ratios are one of a number appear as readmissions when overnight”. Since indicators can of indicators that should be they are on a known treatment be skewed in this way, it is always monitored, and hospital trusts pathway and potentially should in the board’s interest to seek should not rely on a single be excluded from this indicator. assurance from divisional teams indicator. In addition, six factors The calculation then requires the where there is an outlier and, if need to be taken into account use of good statistical methods appropriate, instigate consultant- which can have a direct impact to make the ‘best’ prediction level investigation. on mortality ratios. These are: – but it is possible for two • percentage of the population different answers to be produced, Mortality indicators have taken depending on the methodology. who die in hospital as opposed on a particular importance in to outside the last few years. The three main measures currently used in • population demography ‘A high mortality ratio is England are: not necessarily a sign of • different pathways of care a poor performing trust • SHMI – Summary Hospital-level • zero length of stay emergencies from a safety perspective... Mortality Indicator (Heath & • palliative care a number of factors can Social Care Information Centre) affect an individual figure’ • data quality. • RAMI – Risk Adjusted Mortality Index (CHKS) (For more detail see www.chks.co.uk)

03 briefing 266 The non-executive directors’ guide to hospital data: part four

continued discussion about their Activity data, contracting ‘One of the most effective usefulness as a predictor of the and finance ways to use mortality ratios is safety of care, it is clear that to develop a process of review We have already explained that they are a useful ‘smoke alarm’ around one measure’ the NHS contracting rules set to trigger further investigation. down the principles for which Boards must ensure they have payments should be made. Under an understanding of what the A high mortality ratio is not the standard contract, hospital indicators say about their own necessarily a sign of a poor income is determined by payment organisation and have the performing trust from a safety by results (PbR), which effectively perspective. We have seen a necessary processes in place to is a payment by patients; linked to number of factors which can affect address any issues highlighted. the complexity of the service that an individual figure, including has been provided. the quality of information in the One of the most effective ways case notes to support accurate to use mortality ratios is to This complexity (there are around clinical coding and the service develop a process of review 26,000 codes to describe specific provision, which is reflective around one measure. This will diagnoses and interventions) is of the local population. ensure consistency and will defined by a set of healthcare allow comparison over time resource groups (HRGs). These Non-executives therefore need to to demonstrate improvement. reflect patients who use similar understand what can impact their This review should include amounts of resource. An HRG code own mortality ratio so they can triangulation with other indicators consists of five characters (two identify areas of sub-optimal care, such as infection rates and scores letters followed by two numbers improvements to their clinical from patient surveys. and a final letter) and covers a coding processes, or management spell of care, from admission of patients. Although there is to discharge. At present, there

Figure 1.

Original coding and price Revised coding and price

Diagnosis M65.8 – Synovitis and tenosynovitis M65.8 – Synovitis and tenosynovitis

Diagnosis F17.1 – Harmful use of tobacco F17.1 – Harmful use of tobacco

Procedure W84.6 – Endoscopic excision of synovial plica W84.6 – Endoscopic excision of synovial plica

Procedure Z94.3 – Left sided procedure Z84.6 – Knee site

Procedure Z94.3 – Left sided procedure

HRG (tariff HB99Z – Other procedures for non-trauma – HB23C – Intermediate Knee Procedures for year 13–14) £292 non-trauma without CC £1,673

04 briefing 266 The non-executive directors’ guide to hospital data: part four

are 1,400 HRGs, which are would have resulted in almost used as ‘units of currency’ and £1,400 less income for the trust. ‘Best practice tariffs effectively support standardised healthcare reduce the amount paid for commissioning across the NHS. The tariff is multiplied by a a procedure where there is One patient spell may involve nationally determined market significant unexplained variation several different HRGs, so an forces factor (MFF). This is unique between what is considered automated set of rules is applied to each provider and reflects the good practice and practice by centrally provided software fact that it is more expensive to within the hospital trust’ (known as a ‘grouper’ software) provide services in some parts of to decide the most appropriate the country than in others. database called the Secondary one for payment. In essence, this Uses Service (SUS). Reports from identifies the most costly element Non-executive directors should SUS allow commissioners and of the care provided and assigns be aware that the PbR tariffs providers to make adjustments to the patient spell to that HRG. do change and are used as a monthly contract values agreed lever to control spending in the in the NHS standard contract NHS tariffs are the set prices paid acute sector. For example, best to reflect what has actually for each unit of currency (HRG). practice tariffs have now been happened to patients. For example, £118 is the national introduced. This reduces the tariff for an outpatient attendance amount paid for a procedure In practice it is not unusual for in obstetrics or £5,080 for a hip where there is significant commissioners and acute trusts operation. There are currently unexplained variation between to come to an arrangement over 1,100 tariffs. So the key to what is considered good practice that is different. Under certain getting the right level of payment and practice within the hospital lies in the coding. Many of the trust. The current list includes conditions, commissioners tariffs have two levels: a lower one cholecystectomy, cataract, are allowed to revert to block for the average/normal patient fragility hip fracture care and contracts (a fixed price for all and a higher one for those with acute stroke care, interventional activity irrespective of volumes). complications. The identification radiology, primary total hip and of complications is reliant on knee replacements, adult renal In the NHS Operating Framework coding (and thus on the correct dialysis, transient ischaemic 2010/11 a marginal tariff was information being recorded in the attacks (TIAs), paediatric medicine introduced which meant that source document for coding). and day cases in breast surgery, acute trusts would be paid only general surgery, gynaecology, 30 per cent of the NHS tariff An example of the importance of orthopaedics and urology. price for emergency activity good coding is given in Figure 1, above their 2008/09 levels. This, which shows that the inadvertent When a patient is discharged, together with the non-payment omission of the site being the clinical coders translate the care for readmissions within 30 days knee meant the original coding the patient received into codes of discharge, has continued to the using two classification systems. current day. The money withheld These are ICD-10 for diagnoses from providers in payments for ‘A marginal tariff was introduced and OPCS-4 for interventions (see non-elective admissions above which meant that acute trusts below). This information, together the threshold is administered would be paid only 30 per with other information about the by NHS England’s local area cent of the NHS tariff price patient, such as age and length teams, who use it for “local for emergency activity above of stay, is sent from the hospital’s investment in relevant demand their 2008/09 levels’ computer system to a national management schemes”.

05 briefing 266 The non-executive directors’ guide to hospital data: part four

This is a medical classification list Complications and ‘Benchmarking between trusts by the World Health Organization misadventures can be useful, but too much (WHO). It codes for diseases, emphasis should not be placed signs and symptoms, abnormal Complications and misadventures where rates are higher or lower findings, complaints, social are recorded by the clinical coding that the peer group trust’ circumstances, and external team using the relevant ICD-10 causes of injury or diseases. code for the type of complication. This is taken from the information One further change is that the The code set allows more than documented in the casenotes by calculation of the tariff is changing 14,400 different codes and the clinical team. Benchmarking and will no longer be based on permits the tracking of many between trusts can be useful, but officially calculated costs (reference new diagnoses. This can be too much emphasis should not be costs) but on a less rigid set of expanded to over 16,000 codes placed where rates are higher or information, which will be collected by using further optional sub- lower than the peer group trust. by Monitor (from trusts who classifications. calculate costs at a patient level). A high rate could simply be The Office of Population Censuses down to a trust having more The Department of Health and Surveys Classification of robust information capture and published its last Code of Conduct Interventions and Procedures is not necessarily an indication for Payment by Results for (OPCS-4) is a procedural of poor clinical care. In the 2013/14 in February this year. classification for the coding same way a trust with a lower It is now up to NHS England to of operations, procedures and rate may not be capturing and decide whether further additions interventions performed during recording information very to the code are needed. inpatient stays, day case surgery well and the lower rate may

and some outpatient attendances. disguise a poor performing trust. ICD-10 and OPCS-4 Although the code structure is Trusts should ensure there is a ICD-10 refers to the tenth revision different, as a code set OPCS-4 procedure in place for monitoring of the International Statistical is comparable to the American complications and misadventure Classification of Diseases and Medical Association’s Current cases as part of their clinical Related Health Problems (ICD). Procedural Terminology. governance process.

06 briefing 266 The non-executive directors’ guide to hospital data: part four

The non-executive directors’ guide to hospital data The non-executive directors’ guide to hospital data is a series of Briefings developed by the NHS Confederation and CHKS. The series is intended to increase the non-executive director’s understanding of NHS data and give them the confidence to ask the right questions about it.

All the Briefings are available from the NHS Confederation and CHKS websites.

Hospitals Forum Part one: Activity, pathways and datasets briefing March 2013 Issue 260 This Briefing introduces the scale of NHS activity, the range of activity, the patient pathway and the major datasets. The non-executive directors’ guide to hospital data Part one: Activity, pathways and datasets

Key points Understanding your organisation’s data is an essential part of providing effective oversight. But data may not always give you the complete picture • Detailed data is collected on and it is important to first understand what data is available, how it is service deliveries that take place recorded and what these records are used for. in NHS hospitals in England. This data may be recorded by patient pathway stage, by speciality, by This Briefing will help non-executive directors (NEDs) better understand treatment provided, or by disease NHS data and how it can be used to determine what is going on in their and treatment given. hospital. For the purposes of this Briefing we examine data in the acute care setting only. Data is of course collected in primary care by GPs, • There are a variety of data sources pharmacists, dentists and opticians, but the various datasets are not available to NEDs – including the linked by the NHS. hospital’s patient administration system and the NHS Choose and Book system. This Briefing introduces the scale of NHS activity, the range of activity, the patient pathway and the major datasets. • Analysis of data can provide valuable insights into forecasting maternity services and babies, service demand, and identifying The scale of acute how to better match capacity hospital activity eight million are elective and five to demand by tackling issues million are emergency admissions. like patient non-attendance or In England there are more than 70 appointment rescheduling. million outpatient attendances and Examination of the admissions 15 million hospital admissions a data reveals that the proportion year. An analysis of the total volume of the population admitted to of hospital admissions reveals hospital increases with age; nearly that two million of these relate to two-thirds of patients admitted

Produced in association with

Hospitals Forum Part two: Elective hospital admissions, waiting times and patient experience briefing May 2013 Issue 263 This Briefing looks at elective hospital admissions, waiting times and patient experience.

The non-executive directors’ guide to hospital data Part two: Elective hospital admissions, waiting times and patient experience

Key points Understanding your organisation’s data is an essential part of providing effective oversight. But data may not always give you the complete picture • Elective hospital admissions is an and it is important to first understand what data is available, how it is area where data can be used as an recorded and what these records are used for. indicator of hospital performance. • Length of stay, pre-operative This Briefing will help non-executive directors (NEDs) better understand admissions and variation in NHS data and how it can be used to determine what is going on in their procedure by day of the week hospital. For the purposes of this Briefing we examine data in the acute should be monitored. care setting only. Data is of course collected in primary care by GPs, • Assessing compliance with pharmacists, dentists and opticians, but the various datasets are not waiting times targets is not linked by the NHS. straightforward. The 18-week pathway requires joining primary This Briefing looks at elective hospital admissions, waiting times and care and hospital data sets. Local patient experience. factors also need to be considered. • Patient experience is increasingly Elective hospital admissions in time from the actual admission. used as a quality indicator. The Elective admissions do not include An elective hospital admission is friends and family test is now transfers from another hospital. usually a planned procedure that being implemented and trust has been booked either by the GP scores will be monitored. Elective admissions account for and the patient, or the patient around 50 per cent of admitted has been on a waiting list. An hospital activity in the UK, with elective admission occurs when the emergencies accounting for 35 decision to admit can be separated per cent and maternity and babies

Produced in association with

Hospitals Forum Part three: A&E, non-elective admissions, readmissions and diagnostics briefing May 2013 Issue 263 This Briefing looks at accident and emergency, non-elective admissions, readmissions and diagnostics. The non-executive directors’ guide to hospital data Part two: Elective hospital admissions, waiting times and patient experience

Produced in association with

Hospitals Forum Part four: How to make good use of data – quality and safety, including mortality, briefing June 2013 Issue 266 activity data, contracting and finance. This Briefing looks at how to make good use of data: quality and safety, including mortality, The non-executive directors’ guide to hospital data activity data, contracting and finance. It also includes a short technical guide on ICD-10, Part four: How to make good use of data – quality and safety, including mortality, activity data, contracting and finance

Key points Understanding your organisation’s data is an essential part of providing effective oversight. But data may not always give you the complete picture • Many indicators can be used OPCS-4 and healthcare resource groups (HRGs). and it is important to first understand what data is available, how it is to monitor quality and safety. recorded and what these records are used for. They should not be considered in isolation. This Briefing will help non-executive directors (NEDs) better understand • The quality of indicators is NHS data and how it can be used to determine what is going on in their inextricably linked to the quality hospital. For the purposes of this Briefing we examine data in the acute of data. care setting only. Data is of course collected in primary care by GPs, • The accurate coding of clinical pharmacists, dentists and opticians, but the various datasets are not activity is fundamental, linked by the NHS. particularly when constructing mortality rates. This Briefing looks at how to make good use of data across: quality and safety, including mortality; activity; contracting; and finance. It also • Good clinical coding is also crucial includes a short technical guide on ICD-10, OPCS-4 and healthcare in determining the right level of resource groups (HRGs). payments for healthcare services, and thus defining NHS tariffs. How to make good use non-executive directors need to of data be aware that they have to ask Data, by its nature, is something the right questions in order to that should be handled carefully. establish what the data is telling NHS data is no exception and them. The questions are simple

Produced in association with

07 briefing 266 The non-executive directors’ guide to hospital data: part four

The non-executive directors’ guide to hospital data This Briefing is the fourth in a series of four – the ‘Non-executive directors’ guide to hospital data’ – which have been developed to increase the non-executive director’s understanding of NHS data and give them the confidence to ask the right questions about it. All the Briefings will be available from the NHS Confederation and CHKS websites.

The Hospitals Forum The Hospitals Forum aims to identify the most important issues for hospital service providers, and then work to influence national policy and support and inform members on those priorities. For more information on our work, see www.nhsconfed.org/hospitals or email [email protected] CHKS CHKS, part of Capita plc, is a provider of healthcare intelligence and quality improvement services. In the last 23 years it has worked with 374 healthcare organisations worldwide. For more information, see www.chks.co.uk

Further copies or alternative formats can be requested from: Tel 0870 444 5841 Email [email protected] or visit www.nhsconfed.org/publications The NHS Confederation © The NHS Confederation 2013. You may copy or distribute this work, but you must 50 Broadway London SW1H 0DB give the author credit, you may not use it for commercial purposes, and you may not Tel 020 7799 6666 alter, transform or build upon this work. Email [email protected] Registered Charity no: 1090329 www.nhsconfed.org Stock code: BRI026601 Follow the NHS Confederation on Twitter: @nhsconfed