AHPs Do Count: developing clinical data to inform the service about the work of AHPs

Preface

The series of reports

Acknowledgements

Setting the scene

Purpose of report
Describing AHP Cases

Making Sense of the Health Problem/Issue (HPI)  

Data collected as Free-text
Data collected using International Classification of Functioning, Disability and Health (ICF classification
Data collected using fields from the NCDDP AHP data dictionary

Results

1. Free-text and diagnosis, problems and procedures
2. ICF and its relationship to diagnoses and problems
3. The NCDDP Development Programme

Discussion

Conclusions

Appendix

Preface

The data standards and definitions developed as part of the National Clinical Dataset Development Programme and tested within the AHP census strongly support the AHP Workload report (to be published by the Scottish Executive in September 2006). Supporting AHPs in the use of these data standards in their clinical recording as a way of summarising AHP activity should aid the development of modelling tools for workload and workforce requirements in the future.

This first nationalcensus of AHPs has provided a rich source ofinformation about the numbers of cases being seen by AHPs, access to AHP services, the clinical conditions and the patients' presenting problems as well as the different care objectives of 70% of the patients seen by AHPs on census day (14th September 2005). The response of AHPs in NHS Scotland to this exercise has been overwhelmingly positive in articulating the valuable service they deliver to patients. There are examples in this report of how this information relates to population needs and to the evaluation of service impact.

Back to contents

The series of reports

The previous two reports on the AHP census covered the caseloads of individual disciplines and the putative waiting times and numbers of patients waiting for appointments to see an AHP. Both reports were well received by the AHP community and provided for the first time some quantitative evidence of the diversity, variety and scale of the work done by AHPs.

This is the third report and continues the study of the data collected during the census and concentrates on he clinical data. Each AHP was asked to choose up to 10 interesting cases that they dealt with on census day and provide, for each case, both coded and free-format data describing, for example, the current objective of care and the patient's clinical problems. In all, data on more than 40,000 cases were submitted.

The report AHPs Do Count : relating findings from the AHP census to the AHP electronic community health information project (AHPeCHIP) is published contemporaneously with this report. It is concerned with developing the ideas gleaned from this report and the earlier two reports in order to show how AHPs and related organisations should be tackling the need for and use of information.

All the reports and supporting tables are available on the ISD website and can be accessed through the following hyperlink: http://www.isdscotland.org/allied_health_professionals.

Back to contents

Acknowledgements

The census involved a massive amount of work for already busy AHPs and made heavy demands on the time of the various Heads of Service. It also involved the use of web-based data entry facilities for staff who did not always have easy access to modern computing equipment.

A number of staff from ISD have, at various times and in many ways, made vital contributions in the planning, design, management, analysis and reporting of the census.

Staff from both SEHD and the service community joined with ISD to produce a harmonious working group which demonstrated very clearly the synergies that can be achieved when different skills are combined in a common purpose.

The Project Board offers its heartfelt thanks to everyone involved and who have helped to make this census, the first national census of AHP activity, the landmark success it indubitably is.

We also hope that the work described here will form a baseline upon which AHPs can build useful data models and which in turn will contribute to better patient care.

Back to contents

Setting the scene

Purpose of report

This report is not an exhaustive description or analysis of the clinical data collected in the census; indeed the scale and richness of the data preclude this. Its purpose is to suggest how the data could be analysed and used to inform discussion on what data should in future be collected to reflect and describe AHP activity. Clearly the census data could be analysed in much greater depth, and data extracts can be made available as required to AHPs who wish to use it for their own research studies.

Back to contents

Describing AHP Cases

The analysis of tables and charts in the preliminary report AHPs Count : Preliminary results from the census gave a measure of the different patterns of delivery between professions and demonstrated that simple methods of counting contacts provide relatively little help in understanding the work of the various professions. Instead of focussing on numbers of contacts, it became clear that a description of their caseload (patients or clients for whom an AHP has agreed to provide a duty of care) would provide more meaningful information about AHP activity.

This report demonstrates how free text descriptions of patients' health problems can be classified and how the inclusion of, for example, objectives of care and diagnosis can provide useful information about the work of AHPs.

Back to contents

Making Sense of the Health Problem/Issue (HPI)

The census tackled this in three ways. Firstly, it asked for a description of the patients' presenting problems to be completed in free text. Secondly, AHPs were asked to code the problem at a high level using ICF codes. Thirdly, they were asked to complete a selection of NCDDP codes. Each of these is considered separately below.

Data collected as Free-text

In order to allow the AHPs the freedom to describe their cases in a meaningful way, a 50 character free text field was provided on the data collection instrument in which they could describe the HPI affecting the patient and with which they were working. The clinicians' overall response rate to the census varied from 64% for chiropodists and podiatrists to 94% for dietitians, as detailed in Chart 1 of Preliminary results from the census. [Note that data from therapeutic radiographers was not collected as they had recently been involved in a large UK Cancer Audit].

Free text is not, however, easily converted to statistically analysable data and in order that some analysis of the free text could be made, conversion to qualitative and analysable data was carried out by an AHP researcher. 

The processes undertaken were:

1. A (substantial) list of individual words used within the HPI free-text field was extracted from the census database. Unnecessary words like 'and', 'to', 'with' etc were removed from the list.
2. Obvious spelling errors were corrected, Word linkages (e.g Multiple & Sclerosis) were made and synonyms and abbreviations were expanded.
3. The resulting terms were attributed to the following categories:
   Aetiology - represented a possible cause of a person's HPI.
   Diagnosis - known medical condition either present or suspected.
   Procedure - clinical interventions that had already been carried out usually by someone else or sometimes by the AHP.
   Body Part
   Severity
   Laterality (side of the body)
   Problem - Other reasons for seeing an AHP such as signs, symptoms, functional & social issues.
   Value/Factors - Where information given did not fit into any of the above headings e.g. forensic status, child, baby, elderly person.

As an example, the table below shows how three HPIs each from a different profession were classified. The relative complexities are indicated by the number of categories to which they have been assigned. Thus the text could be coded using both the medically orientated ICD10/OPCS4 classifications and the functional ICF classification. AHPs do not regularly use these classifications and need to be encouraged to standardise their clinical descriptions by using them.

Table A

Back to contents

Data collected using International Classification of Functioning, Disability and Health (ICF) classification

AHPs were asked to categorise the main problem or issue that they would be dealing with using the ICF classification (using a high-level classification only).

The aims of ICF are to

• provide a scientific basis for understanding and studying health and health-related states, out comes and determinants
• establish a common language for describing health and health related states in order to improve communication between different users, such as health care workers, researchers, policy-makers and the public including people with disabilities
• permit comparison of data across countries, health care disciplines, services and time
• provide a systematic coding scheme for health information systems.

This classification focuses on the "components of health" as opposed to more traditional "consequences of disease". Very often AHPs are dealing with patients' problems rather than the disease itself. ICF is designed to measure and detail impairments and limitations of problems and its categories can be grouped to various high level headings.

Within the AHP electronic community health information project (eCHIP), AHPs got together within discipline to discuss common terms that described the clinical cases with which they were involved. There was no common agreement that any of the established medical coding structures could be used universally across all nine professions to describe their activity or measure the resource they dedicated to patient care. Some work had been done in other countries in utilising ICF in some of the professions to describe the changes in health states (health outcome) through therapeutic intervention.

Details of the data requested and the subsequent analyses are discussed in section 2 of the results below. 

Back to contents

Data collected using fields from the NCDDP AHP data dictionary

AHPs were asked to complete a number of fields all taken from the draft AHP NCDDP data dictionary. This allowed the fields to be tested to see whether AHPs could understand and complete them.

The National Clinical Dataset Development Programme (NCDDP) supports clinicians in developing interoperable national data standards in order to facilitate the implementation of integrated care records across NHS Scotland and an AHP has been seconded to the programme to help develop the AHP standards and dictionary. The dictionary consists of a number of defined standards agreed by the AHP community to be useful in describing activity; full details can be found at http://www.datadictionary.scot.nhs.uk/

Back to contents

Results

This report looks at three main features of the data:
1. The relationships between diagnoses, procedures and problems.
2. ICF and its relationship to diagnoses and problems
3. The use of NCDDP fields.
Note that the findings only relate to the particular sample of patient episodes that each individual clinician chose to describe. This means that the percentagesin the different clinicalcategories cannot be extrapolated to give percentages in these categories for all patient episodes.

1. Free-text and diagnosis, problems and procedures.

The process of categorising the free text used to describe the Health Problem/Issue (HPI) for each case described resulted, after removal of duplicates etc, in a total of free text strings each describing the HPI for 39,877 cases. This number is used as the denominator for the proportions (expressed as percentages) in the Venn diagrams.

Some 95% of the text strings contained a word or words which could be classified as a Diagnosis or Aetiology, a Problem or a Procedure.

The Venn diagrams themselves show the variationsin the relative proportions of these three categories. Each of the three circles represents (although not proportionally) one of Diagnoses and Aetiologies, Problems, and Procedures with the intersections representing those cases where the HPI was assigned to two classifications or in a small number of instances (1% of the total), all three.

Venn Diagram 1 shows for all disciplines the numbers of HPIs assigned to each of the three categories and combinations of categories. The associated table shows the number of HPIs in each sector of the diagram.

Venn Diagram 1

Altogether 63% (two thirds) of the HPI text strings contained a word or words that could be categorised into the group 'Diagnoses and Aetiologies'. The diagram shows that of these 39% could only be categorised to this group. The remaining 24% included a word or words that could also be categorised as a problem or as a procedure or as both. Similarly 48% of the HPI text strings included words that could be categorised as a problem and only 15% as a procedure.

Venn diagram 1 shows that AHPs as a group tended to describe the patients' HPIsin terms of diagnoses rather than as problems. A comparatively small proportion described them in terms of clinical procedures.

Venn diagrams 2-11 show for each profession the distribution of the three categories. The table beside each diagram shows the top ten words (diagnosis/aetiology, problem or procedure) used by the profession in describing their patients. The table below shows, by profession, the number of HPIs described in the Venn diagrams.

Table B
 

Venn Diagram 2

Venn Diagram 3

Venn Diagram 4

Venn Diagram 5

Venn Diagram 6

Venn Diagram 7

Venn Diagram 8

Venn Diagram 9

Venn Diagram 10

Venn Diagram 11

For most of the professions the top ten cover about 50% of the words used to describe the categories. The main exceptions are Arts Therapists with 65%, Podiatry with 67% and Prosthetists with 99%. The figure for the Prosthetists is remarkable particularly since the figure for Orthotists is 47% suggesting a different approach to the census and possibly their patients (This finding is repeated in different circumstances later in this report). It also reflects the greater diversity of conditions seen by different professional groups. 

It is also interesting to note that there is very little overlap between professions in the words used. With the exceptions of Diabetes, Arthritis, Stroke, Fractures, Pain and Assessment most of the top ten words are unique to a profession.

These figures do not reflect entirely the diversity of work undertaken by the AHPs since they were asked specifically to include up to ten representative cases from the total number of cases they saw on that day.

Back to contents

2. ICF and its relationship to diagnoses and problems

For each case chosen for inclusion in the census the AHP was asked to classify the patient's health problem
or issue (HPI) using the International Classification of Function, Disability and Health code (ICF).

The classification is divided into 4 main groups: 

• Body Structure (BS)
• Activities and participation (AP)
• Body Function (BF)
• Environmental factors (EF)

In this census the AHP was only allowed to choose one code and so, at the outset and with little training or instruction, had to choose which of these top-level groups was the correct one to classify the patient's HPI. It is recognised that for some HPIs different AHPs may have made a different choice of group. Improved consistency will be achieved through greater training in the use of ICF and clinical coding which will enable more thorough comparative analysis within and between professions.

For the sample of patients that AHPs elected to describe, over 90% of the HPIs could be classified using the ICF codes. Chart 1 below shows the relative proportions of each group for each profession for the selected cases.

Chart 1

The classification groupings used can be related in very general terms to the model of care delivery used by a profession or by substantial groups within a profession. These vary across a wide spectrum from a highly clinical model for Prosthetists working with patients who have suffered loss of body structures at one end, to Speech & Language therapists who work in a much more facilitative and enabling role with people who have problems in the area of Activity and Participation. In between are professions like physiotherapy, covering many different clinical conditions and care groups with a range of models from direct clinical care to facilitation and enablement. In chart 1 the professions are arranged such that the possible social model decreases from left to right whilst the medical model increases from left to right. The order of the professions in the chart is perhaps what might be expected, with speech and language therapists seeing the problems from the point of view of relating to everyday life and prosthetists seeing the problem very much in relation to body structure. An interesting comparison is between prosthetists and orthotists; whilst prosthetists use Body Structure codes 100% of the time the orthotists use them only 40% of the time and use the Activity and Participation codes for 5%. This suggests that whilst they do (physically) much the same type of work their view of what they are doing differs and may reflect a very different approach to their patients.

A small number of clinicians, less than 10 out of 6,000, found high level mapping not detailed enough, for example swallowing problems were difficult to classify from the high level codes only. Despite this and noting the 2% of cases where the AHPs were unable to match to ICF, the Census has demonstrated that the ICF classifications can also be used to compare cases across the AHP disciplines. In the future at a high level it could also be used to note trends in clinical practice and could be used as an outcome measure. ICF has now been included in NCDDP data standards.

As described above under 'Making Sense of the Health Problem/Issue (HPI)' the free text provided by each AHP was parsed into a number of categories representing diagnosis, problems, aetiology etc. It is interesting to speculate how the major ideas from this text relate to the resultant ICF codes.

The table below shows that some diagnoses were amongst the top ten for a number of professions.

Table 1

Taking the diagnoses Arthritis and Dementia as examples, Charts 2 and 3 below shows the relative percentages each discipline has allocated to the main ICF groupings of body function, body structure, activities and participation, and environmental.

Chart 2

Chart 3

In arthritis most of the disciplines have tended to classify it to body structures or body functions; in dementia the classification is to body functions then structure. The use of activities and participation is, in the main, left to the occupational therapists. Again the strong medical model is noted. 

Both of these diagnoses are seen and treated by occupational therapists (amongst others). While the underlying illnesses are quite different in cause and in treatment both show quite strong similarities in presenting symptoms, for example they both demonstrate problems with general daily tasks, self care, domestic life and, to some extent, movement functions and it is with these symptoms that the occupational therapists would be involved. The treatments provided themselves will not, of course, necessarily be the same. It might therefore be expected that the classification to the ICF groupings would be similar. However, in Arthritis the allocation is more to body structures than functions or activities whilst in Dementia body functions are stressed with almost no allocation to activities.

As was shown in the Venn diagrams above, the AHPs also identified problems as well as diagnoses. Chart 4 below shows how the identified problem 'pain' is distributed amongst the disciplines in terms of the ICF groupings.

Chart 4

Again with the exception of the occupational therapists the majority of AHPs identifying this problem allocated it to the Body Structures group.

The International Classification of Disease - 10th revision (ICD10) provides an aetiological framework for the classification of health conditions (diseases, disorders, injuries etc) and this overlaps in part with the ICF classification. Impairments as signs and symptoms are part of a constellation that forms a 'disease' (ICD10)whereas ICF uses impairments as problems of body functions and structures associated with health conditions. Through the analysis of the HPI within the census it has become clear that where AHPs use ICF to describe the patient's health problem they must also provide information about the aetiology of the problem. For example many cases of hemiplegia were reported, but unless the aetiology is provided we do not know if the hemiplegia was associated with a Head Injury, Stroke, Cerebral Palsy, Cancer, Spinal Injury etc.

Back to contents

3. The NCDDP Development Programme

The National Clinical Dataset Development Programme supports clinicians in developing interoperable national data standards to facilitate the implementation of integrated care records across NHS Scotland.

The standards will:

• support direct patient care, being informed by prevailing best practice guidance
• facilitate effective communication between health care professionals
• improve data quality and support secondary data requirements where possible including data to support clinical governance
• be freely and widely available through publication in the web based Health & Social Care Data Dictionary
• incorporate agreed national clinical definitions and implement national terminology
• be UK compatible where possible.

A National Consultation on the AHP Core Data Standards ran simultaneously with the Census and contained the following data items which were also used in the AHP Census:

• Health Problem/Issues
• Patient Journey Stage,
• Principal Objective of Care and
• Caseload Status.

The census results for two of the NCDDP items - objective of care and patient journey are discussed in this section of the report.

Charts 5-8 below show the distribution of objective of care for each of the defined stages of the patient journey.

Chart 5

Chart 6

Chart 7

Chart 8

These charts demonstrate that some AHPs seem to have misunderstood what the stages in the patient journey mean, for example, chart 5 with stage Prevention has a relatively high proportion of objective of care other than 'prevent'. For the other patient journey stages the types of objectives of care used are very limited and those that have been are very close in meaning to each other. For example for patient journey stage 'Living with' the types of objective of care used were enablement/facilitation and maximise potential and they cover 68% of the returns. This implies that the patient journey item is contributing little to an increased understanding of the pathway of care. This evidence supports the NCDDP's decision to drop the patient journey as an item in the dictionary. 

Back to contents

Discussion

The need for standardisation of recording of clinical information is well highlighted by the results above. This is well-known and has been reported in many previous publications (Appendix A). Generally the purpose of clinical data standards is to create consistency and to facilitate:

• Care: Service delivery through support for direct patient care by collecting and having available meaningful clinical information.
• Share: Sharing of patient information between HCPs which is clear, unambiguous and of high quality with the data being collected only once.
• Measure: Support measurement of performance indicators to allow comparisons and for commissioning of servicesas a by-product of direct clinical care.
• Compare: Monitor conformance to service e standards, benchmarks or targets at a local or national level in support of continuous quality improvement.
• Prevent: Inconclusive results in research trials and in meta-analysis of research.

To date, much of the nationally consistent data collected in NHS Scotland has, in the main, been through central returns to ISD relating to patient administrative and demographic data and NHS hospital activity data. This is reflected in the content of the SMR Data Manual and the Definitions and Codes Manual. AHPs have on the whole not engaged in this process and are therefore less familiar with data standards although it is hoped that the introduction of the AHP NCDDP data dictionary will change this.

Prior to the AHP census being undertaken in September 2005 limited national information was available on the work of AHPs. Nationally available information on all AHP services is currently limited to quarterly aggregate counts of 'new patients' and 'total attendances', collected as part of ISD(S)1 scheme. Although collected for many years, this data has been of variable quality, thus reducing the utility and value of the information.

Historically, a range of factors have affected the quality of AHP data on ISD(S)1. Difficulties in providing suitable national definitions and recording guidance led to local differences in interpretation. In some areas, for some professions, there have been problems in capturing the required information. In the absence of any robust national quality assurance of ISD(S)1 AHP data, anecdotal reports fuelled the uncertainty regarding data quality.

In addition to ISD(S)1, there have been a variety of data collection exercises over the years, relatedin varying degrees to particular professions, including: chiropody services, general ophthalmic services, diagnostic radiology/radiography workload statistics, the UK amputee statisticaldatabase, podiatryinformation project, speech and language therapy statistics.

Importantly, none of these, including ISD(S)1, are considered to record sufficient and/or appropriate information on all or indeed any AHP service. In the case of ISD(S)1, the aggregate counts of contacts are of limited use, particularly for those professions whose patients are commonly characterised as having conditions which require ongoing care, and the definitions are not themselves always consistent. Also, aggregated data sets such as ISD(S)1 do not provide any information on the range of services provided by AHPs; there is no information on, for example, AHP caseloads, patient health problems, patient journeys, objectives of care or waiting times. 

Back to contents

Conclusions

Appendix

1. NHSScotland. (2004) National eHealth/ IM&T Strategy 2004 - 2008

2. SEHD (2004). AHP Research & Development Action Plan 2004

4. Hicks N. (1997) Evidence based health care. Bandolier 4: 39, 8.

Main contact: Jenni Muir

 Printer friendly version