Herd health plans (HHPs) must be carried out on an annual basis for the majority of dairy farms. In the UK, all dairies that are Red Tractor-accredited are required to have an annual health and performance review, as part of the HHP (Red Tractor, 2022).

When making farm-specific recommendations, it is important to have an understanding of the current level of performance and disease on your farms.

Data analysis is, therefore, an essential diagnostic tool when making herd health recommendations, and it is also important for monitoring progress.

Ultimately, if compliance with the herd health plan is poor, data analysis can be used to demonstrate a lack of progress, or even worsening of performance.

This article will review the different sources of data available for vets and advisors, and give tips for making the most out of farm data.

Understanding different sources of data

Advisors should be aware of the types of data that may be available on farm. Broadly, this can be split into data held on farm (for example, in a farm diary or on-farm software), or data held remote from the farm (for example, the vet practice, technical services, nutritionist, consultant, milk recording organisation, milk buyer or genetic database).

Data held on farm is usually more accurate; however, it relies on a member of farm staff to record it. At a basic level, all events and health treatments should be recorded in a farm diary; however, analysing paper records can be time consuming. Information recorded in an on-farm software is usually more consistently recorded and accessible. Remote data is usually a very consistent method, as data are recorded systematically. However, remote data sources only record what is relevant to the company collecting the information; for example, milk recording organisations usually record no information on youngstock, which can limit fertility analysis. 

Before analysing data, it is important to understand what is being recorded. Table 1 gives examples of the events that may be useful for different areas of analysis. Broadly speaking, the more information a farmer is recording, the more useful it can be for herd health planning.

In farms where natural mating is used, service dates might not be recorded, which can limit fertility analysis.

Objectives and motivation

No sense exists in analysing data without a clear objective – advisors should have a working hypothesis; that is, what are you trying to show? This may be influenced by farmer objectives and motivation.

Objectives may be short-term or long-term, they may be aspirational, or may be a requirement of their contracts. Aspirational goals are aiming towards constant improvement; farmers are simply trying to get better. For contract-related goals, farmers are often working towards a defined target, and this may be linked to payment incentives or penalties.

The more that farmers engage with their targets, the more they are likely to achieve them – longer term, sustained engagement is important to maintain progress (Rose et al, 2018).

One proposed model for farmer motivation is summarised in Figure 1, adapted from Ritter et al (2016).

Figure 1. A proposed model for classifying farmer motivation, adapted from Ritter et al (2016).

This model splits farmers into four groups based on their perception of the importance of a problem, and their belief in their advisor(s) to solve the problem. In the study by Ritter et al, the model is used to demonstrate approaches to Johne’s disease control, but it can be applied to other areas of herd health.

1. Proactive farmers believe they have a problem, and are eager to make changes based on advice. They will often approach their vet or advisor, may have already implemented your advice, and are seeing the benefits.

2. Disillusioned farmers acknowledge they have a disease problem, but do not believe in the proposed strategies. These farmers need to be convinced that your advice is sound; they may benefit from case studies or farmer discussion groups (particularly including proactive farmers).

3. Unconcerned farmers believe in your advice, but do not accept that they have a problem. These farmers would benefit from data analysis, benchmarking, or a cost calculator to highlight their problems.

4. Deniers are those who do not believe they have a problem, and do not believe in your advice. These are the most difficult group to work with. These farmers often respond better to the “stick” approach, where targets are enforced on them either through legislation or their milk buyer. Eventually, “deniers” will be brought along once the unconcerned and disillusioned are converted to proactive.

It should be stressed that this is just one model of thinking about farmer motivation. Whatever approach you use as an advisor, understanding farmer motivation is key to improving herd health.

Targets and language

When writing HHPs, it is common to include targets. To set effective targets, it is essential to have an underlying understanding of the farming system and the associated language.

All-year round calving herds use different fertility language to block calvers (see Table 2), and failure to adopt the correct terminology can reduce the impact of your advice. Language used around mastitis can also be unclear – it is useful to clearly define the difference between a clinical mastitis case (that is, clots or blood present in the milk) and subclinical mastitis (that is, elevated somatic cell count without visible changes). 

When setting targets, it is important they are clearly defined, realistic and achievable. Vets and advisors should be familiar with any contractual targets that may exist for farmers on aligned contracts. The way that contractual targets are set can make a difference to farmer motivation: one study found that farmers were more motivated by penalties for high cell count milk than an equivalent bonus for low cell count milk (Valeeva et al, 2007).

Key performance indicators (KPIs) are single statistics that can be used to measure performance in one area. Several publications use KPIs to monitor performance of the UK dairy industry over time, and readers are directed towards these useful reports (Hanks et al, 2024; Kingshay, 2023; Leach et al, 2024).

Benchmarking farm performance against published figures can be a useful tool to stimulate farmer discussion and motivate farmers to improve relative to their peers (Sumner et al, 2018). This is particularly useful for the “unconcerned” group of farmers in the study by Ritter et al (2016). Farmers can be benchmarked against other farms in the country, within the milk buyer pool or within your group of clients.

Cost calculators can be useful for farmers more motivated by economics. Figure 2 shows an example of a mastitis cost calculator that is available to vets and advisors through QuarterPRO (Quality Milk Management Services, 2021).

Figure 2. An example of the costs of mastitis in a 200-cow herd. Cost calculator available for free at tinyurl.com/bdeevxkt

On top of the visible costs (drugs, labour, milk discard), it is important to highlight “hidden” costs that farmers may be less aware of. In this example the biggest costs of mastitis are hidden – reduced milk production and increased risk of culling.

Other farmers may be motivated by things that are less easily measured; for example, better welfare. Vets and advisors can monitor “proxies”, figures used to represent a less easily measured outcome, often a subjective or objective measure of animal behaviour; for example, Cow Comfort Index, defined as the proportion of cows in contact with a cubicle that are standing, is a reasonable proxy for cubicle suitability (Cook et al, 2005).

The proportion of cows exiting the herd in the first 100 days in milk is a reasonable proxy for “unplanned exits”; that is, those caused by early lactation disease or injury (Pinedo et al, 2010). 

Use of tools – data analysis software

The number of data capture software and data formats is increasing in all industries. To keep up with farmer demand, vets and advisors need to be familiar with data analysis software.

Basic calculations can be carried out in spreadsheets, but more advanced assessment of farm performance requires bespoke software. This software can help identify the limitations of farmer data. Most software will produce a list of the kind of event data highlighted in Table 1, and this is a useful place to start. Where farms are not recording everything they could be, more regular herd health advice can help to show value and encourage better record keeping.

Farmers should be shown the value of better recording and given clear advice about what, and how, to record better.

Figure 3. Survival analysis of time to first service and first conception for lactating cows within 300 days.

Most dairy farmers are interested in fertility performance. Figure 3 shows a survival analysis of time to first service and first conception, as an example of one of the tools that can be used to generate discussion on farm.

1. The farm is not sticking to a voluntary wait period of 42 days; 20% of cows are served earlier in lactation, though few of these serves result in a successful pregnancy.

2. The majority of cows are served by 100 days in milk; some cows are served for the first time between 100 to 200 days in milk, but few are served for the first time after more than 200 days.

3. Conception during the first 100 days in milk is low and improves between 100 to 200 days in milk

4. 16% of cows are never served in a lactation cycle – this is a proxy for non-fertility culls or herd exits; that is, cows deliberately not served or exiting the herd prior to service.

5. 10% of cows are served, but do not conceive by 300 days in milk – this is a proxy for fertility culls; that is, cows the farmer wanted to get pregnant, but failed to conceive.

Going further and looking to the future

In the UK, several structured approaches to managing dairy cow health exist. Examples include the Mastitis Control Plan, QuarterPRO (Agriculture and Horticulture Development Board), Healthy Feet Programme, National Johne’s Management Plan, Cattle Health and Certification Standards-approved health schemes, BVDFree (England) and Gwaredu BVD (Wales). All of these programmes require a level of diagnostic testing and data analysis to identify patterns of disease leading to farm-specific advice.

These structured approaches are ideal for progressive farmers, but can also be useful for those disillusioned with previous advice or unconcerned about their current health status.

Use of technology in farming is accelerating, which means more data are likely to be available to farmers, vets and advisors. Several new and emerging technologies will soon be generating data for use in herd health. Some examples include:

• Camera technology to identify lameness, body condition score, assess cow behaviour and welfare assessment.
• Automated mastitis detection in conventional and automatic milking systems.
• Machine learning tools that predict and model production, health and efficiency.
• Improved integration of different monitoring systems.

Conclusions

• Use of data in every industry, including farming, is growing.
• Data analysis is a key skill for modern dairy vets and advisors.
• Robust data analysis encourages better farmer engagement, improved recording and, ultimately, more fee-paying consultancy.
• Data analysis software improves the speed and reliability of analysis.
• It is important to understand your client’s objectives, motivations and language when giving herd health advice.