|.......||Poultry Health Monitoring - Benefits
of Computerised Systems.
Paul McMullin, Poultry Health Services
Presented at a Worlds Poultry Science Association Symposium : Transfer of Knowledge in Poultry Production
17-19 November 1997 Utrecht
Generation of knowledge with respect to poultry health usually involves 2 stages: the collection of data (a farm visit, post-mortem examination, serology test, etc), and its analysis. Transfer of knowledge is usually by the preparation of a written report and the sending of the report to the right person. Manual systems fulfilling the requirements of these stages for the generation and transfer of knowledge have been in use in most, if not all, poultry veterinary practices for many years. Inexpensive yet powerful micro-computers have made it possible to automate these manual systems. The cost and effort is often repaid simply by improvements in administrative efficiency. The ability to easily summarise and analyse an accumulated data-base of information then becomes a 'free' bonus. However the mere fact that it is possible to automate an existing manual system does not mean that this is always the best approach to automation. Some years ago we initiated the process of implementing a computerised system of managing the data of a busy poultry veterinary practice. At that time no off-the-shelf software was available to fulfil our requirements. A modular approach was used in which separate systems were gradually implemented to cover each area: post-mortem examinations, serology, 'field' bacteriology (hygiene on-farm and in hatcheries, Salmonella, Campylobacter etc.), quantitative microbiology (mainly for food hygiene), prescribed soluble medicines and vaccines, in-feed medication ('Medicated Feedingstuffs Prescriptions'). The system for dealing with post-mortem examinations is described in some detail as an example of the structure of all of these systems. Serological data related to the detection and monitoring of response to Salmonella vaccination, is also used to illustrate the benefits of this approach.
Computers have been used for the collection and analysis of veterinary data for at least 25 years (Appleb et alii 1971). In recent years the fall in cost of the equipment has permitted much more widespread use. New veterinary information systems continue to be developed (Hanks et alii 1994). The general principles of veterinary information management have been covered elsewhere (Thrusfield, 1983, Blood and Brightling, 1991). This article deals with the application of these principles to the data processing requirements of a poultry veterinary practice.
Generation and transfer of knowledge with respect to poultry health can be very simple. The generation of knowledge involves 2 stages: the collection of data (a farm visit, post-mortem examination, serology test, etc) and its analysis. Transfer of knowledge may be through the preparation of a written report and the sending of the report to the right person, but may also be carried out face-to-face in meetings with farm staff and senior managers, or in training courses. Manual systems fulfilling the requirements of these stages have been in use in most, if not all, poultry veterinary practices for many years. There are only 2 disadvantages of such systems :
1. They can absorb much secretarial time simply transcribing information, and veterinary time to ensure that transcription errors have not occurred.
2. Analysis of the data generated over time, by farm, company, medicine, etc., is very difficult, labour intensive, and, hence, will only be done in exceptional circumstances.
Figure 1 below shows the flow of information for a typical manual system to record and report the results of post-mortem examinations. The simplest approach is to convert the essence of an existing manual system into a computerised data-base which will avoid the need to transcribe data. This may not always be the best approach however. The advantage of doing so is that staff and clients will find it easier to adapt to the change. The disadvantages are that the particular manual system may not suit itself to computerising or take advantage of the power of the computer, and that, in most cases, a set of custom programmes will need to be written}
Figure 2 below shows the information flow in a computerised version of the same system shown in figure 1. In judging the cost of implementing such systems it is important to take into account the cost of the equipment (hardware), the cost of 'off the shelf' software, as well as the cost of any programming and staff training. Where custom programming by third parties is involved an extra 'hidden' cost will be the amount of veterinary and administrative staff time taken up with specifying the system and educating the programmers about the specialised nature of poultry veterinary medicine.
The cost and effort is often repaid simply by improvements in administrative efficiency. The ability to easily summarise and analyse an accumulated data-base of information then becomes a 'free' bonus.
Some years ago we initiated the process of implementing a computerised system of managing the data of a busy poultry veterinary practice. At that time no off-the-shelf software was available to fulfil our requirements. Rather than attempt to complete the process in one operation a modular approach was used. Separate systems were gradually implemented to cover each area: post-mortem examinations, serology, 'field' bacteriology (hygiene on-farm and in hatcheries, Salmonella, Campylobacter etc.), quantitative microbiology (mainly for food hygiene), prescribed soluble medicines and vaccines, in-feed medication ('Medicated Feedingstuffs Prescriptions'). Though based on pre-existing manual techniques, these systems have developed to take advantage of the available computing power.
PostMort - A typical veterinary data reporting system.
A large amount of data relevant to the pathological processes associated with endemic disease is generated from routine post-mortem examinations of material submitted from affected flocks. Traditional manual and computerised systems for handling these data tend to be based on free-form text and are aimed at the generation of a report of the examination. The objective of this project was to develop a system which would facilitate routine administration and filing but which would also allow long term accumulation, tracking and analysis of data. \par \par Since many post-mortem examinations are carried out on material which is sent to the veterinary practice, the first thing to consider was the transfer of knowledge from the farm to the veterinarian. Different sample submission forms were reviewed, and the form shown in Figure 3 was the result. The top half is the front of the form and provides space for the farm to provide detailed information. The bottom half is actually the reverse of the form and contains 3 sections:
1. Instructions and hints on the taking and submission of samples
2. An address label pre-printed with the name and address of the veterinary practice. By folding the form in 4 and placing it in a self-adhesive clear plastic 'documents enclosed' pouch this may be used as a label.
3. The bottom section for use in recording the results by the veterinarian or practice staff.
The same form may be used for the collection of data during farm visits and when field post-mortems are carried out. A prototype system was developed with a simple commercial data-base programme. This was used to define appropriate data structures. It quickly became evident that such a system could relatively easily be adapted to provide usable reports. However, the consistency of retrieval of data for long-term analysis was poor. This was due to subtle differences in the way the data were recorded. The answer was to develop coding systems for species, company, farm, and pathological findings. To facilitate this the prototype system was implemented and improved as a series of programmes written in QuickBasic. The systems would automatically check the codes entered against those recognised by the system and ask for confirmation if they were unrecognised. The coding system for 'finding' allows both pathological coding using site and type of lesion (i.e. ASI for Air Sac Inflammation) and also codes for syndromes (ABD for Acute Bursal Disease). Figure 4 shows the screen used to record up to 5 findings for up to 14 birds or animals. As each finding code is entered its interpretation is shown at the right hand side. Note that it matches the data-recording area on the submission form (Fig.4).
Figure 4. The Submission Form for post-mortem material. Forms used for other types of sample are in the same style but adapted to the nature of the work and the relevant computer system.
Request for Post-Mortem Examination ====.=========.=====.==|= Lab Reference: P-= -==
Vaccination and Medication History
Signs noted (tick one or more): ========================..=Mortality ...Culls
For Laboratory Use Only:
SEX =1===2== 3==
Figure 4. Screen for entering post-mortem results
As soon as the post-mortem findings are entered the system can present a summary report for the group of birds. It sorts the individual birds from heaviest to lightest and sorts the findings from most common to least common. These are then printed in a grid in which it is readily appreciable which findings tend to be associated with one another and any association with weight, sex or type (dead or cull) of bird. Such a grid is illustrated in Figure 5 below. Note that although the finding codes are used in the grid, the full translation of each code (with number of birds affected) is shown in detail in a 'Diagnoses' section.
Figure 5. The PM findings section of a report.
A sample of 6 broiler chicken carcasses was submitted because of rising mortality (18/33/51).
Routine reports may include sections for laboratory results (bacteriology, sensitivity testing, histology, parasitology), and will usually end with comments and suggested actions. There remains only to print the report and ensure that it is sent to the correct person.
The system looks up the appropriate reportee(s) based on the combination of the code used for the company and the code used for the species. It uses this information to print the name and address at the top of the report (and print multiple reports if there are multiple reportees) as is shown in Figure 6. The positioning of the name and address is intended to place this information in the \'93window\'94 of standard windowed envelopes when the reports are folded for dispatch.
Figure 6 Report Header
Graphs may also be produced over time or simply showing all birds with a particular finding, weight against age (as in Figure 8 below) . By comparing the distribution against the normal growth profile it is possible to establish whether the affected birds have been growing above or below standard.
Figure 8. Graph of turkeys with oesophageal ulceration.
The same basic principles have been applied to the other systems we use, with modifications appropriate the requirements of each system. The serology system, for instance, provides both for simple data-entry for recording HI and agglutination tests and for direct reading of results from an automated Elisa plate reader. Because of the much greater complexity in managing the work load (differing sets of assays, numbers of sera to be tested, numbers of houses examined etc.) it also includes detailed functions for monitoring the pattern or work and assays to do. Like the post-mortem system it facilitates the generation of reports in a standard fashion, provides for detailed listings and analyses, and allows easy importation of individual sera data and group means into spread-sheet programmes
The graph shown in Figure 9 was prepared in order to examine the response to Salmonella enteritidis vaccination in broiler parent chickens. All birds were vaccinated with 2 doses of an inactivated product (Salenvac, Hoechst) at about 11 and 16 weeks of age (McMullin et alii 1997). The data for 156 groups of sera from vaccinated birds were extracted. A peak response was found at 22 weeks of age with a gradual decline in lay. Regression lines were calculated for the period of immunisation and the period of decline. This sort of approach provides a simple approach to establishing a 'normal' response pattern against which future results may be readily compared. We have prepared similar charts for most assays used in our laboratory to evaluate response to vaccination (McMullin 1995).
Figure 9. Mean S/P Band in Elisa Assay (Guildhay) for groups of sera from broiler parent chickens vaccinated against Salmonella enteritidis. Each point represents the mean of 5 to 20 birds.
Computers can greatly facilitate the routine handling of poultry health information generated in a veterinary practice. The main benefits lie in avoiding transcription and speeding-up report generation. In order to get the full benefit of this it may be necessary to change the way we do thing. It helps, for instance, if the structure of sample submission forms and any data-recording forms match the equivalent data-entry screen on the computer. The development of easily-remembered codes for farms, PM findings, companies, assays etc, both speeds up data-entry and also improves its accuracy, especially when the system checks the codes against a data-base of correct codes. If effort is devoted to ensuring the accuracy of the data recorded, the real benefit will be the accumulation of an historical data-base of information from which trends of disease occurrence can compared to intervention strategies. The next challenge will be to develop systems which will allow the incorporation of data relevant to health control strategies from various sources (veterinary practice, processing plant, productivity parameters).
Appleby, EC, A.H.S. Hayward and R. Renyard (1971) The use of a computerised record system in the analysis of 391 cases of suspected neoplasia. The Veterinary Record 89:547-552
Blood, D.C. and P.Brightling (1991) Veterinary Information Management. Balliere Tindall
Hanks, J.D., B.G. Bedard, Samkhan Navis, Budi Tri Akosos, S.N.H.Putt, A.D.James and Agus Heriyanto (1994). DIAG, a laboratory information management system developed for regional disease diagnostic laboratories in Indonesia. Tropical Animal Health and Production 26:13-19
McMullin, P.F. (1995) The Organisation and Use of a Serological Monitoring Service for the Poultry Industry. Dissertation for the Diploma in Poultry Medicine and Production. Royal College of Veterinary Surgeons (lodged in Wellcome Library, RCVS, Belgravia House, 62-64 Horseferry Rd London SW1P 2AF).
McMullin, P.F. (1997) The development and application of a micro-computer data-base system for the collection, administration and analysis of poultry post-mortem data.World Veterinary Poultry Association Congress, Budapest. Abstract No. 242 Poster presentation.
McMullin, P.F. & R. Higgins (1997) Idiopathic ulcers of the palate, pharynx, and distal oesophagus in commercial growing turkeys:Gross and microscopic lesions and patterns of occurrence. World Veterinary Poultry Association Congress, Budapest. Abstract No.361. Poster presentation.
McMullin, P.F, K.R.Gooderham and G. Hayes (1997) A commercial Salmonella enteritidis ELISA test: Results arising from its use in monitoring for infection and response to inactivated vaccine. World Veterinary Poultry Association Congress, Budapest. Full paper in proceedings of EU COST sessions.
Thrusfield, M.V. (1983) Application of computer technology to the collection, analysis and use of veterinary data. The Veterinary Record 112:538-543 o