Welfare Assessment in Non-Human Primates Used in Research

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• Welfare assessment in non-human primates (NHPs) necessitates a multimodal approach integrating behavioral observation, physiological measurements (e.g., acute phase proteins, fecal glucocorticoid metabolites), and clinical judgment to distinguish adaptive responses from sustained welfare compromise.
• Species-typical behaviors (e.g., locomotion, social grooming, foraging) serve as crucial baselines; deviations in frequency, duration, or context, alongside the presence of abnormal repetitive behaviors (stereotypies), signal potential welfare concerns.
• Individual variation in personality and coping style significantly influences indicator interpretation; establishing individual baselines and considering proactive versus reactive coping mechanisms is essential for accurate assessment.
• Structured welfare assessment protocols with defined observation windows, operationally defined scoring scales, and established inter-observer reliability are critical for repeatable and reliable data collection, guiding intervention thresholds.
• Physiological indicators like acute phase proteins can detect subclinical inflammation or stress responses, while fecal glucocorticoid metabolites offer insights into chronic stress, but interpretation requires species-specific reference intervals and consideration of confounding factors like sampling stress.
• Documentation of longitudinal tracking, including baseline assessments and all subsequent scores and interventions, is vital for individual animal care decisions and facility-level welfare monitoring, enabling trend analysis and identification of management issues.

Welfare assessment in non-human primates (NHPs) used in research requires a structured, evidence-based approach that integrates behavioral observation, physiological measurement, and clinical judgment. This article provides veterinary researchers and laboratory animal clinicians with a framework for selecting, implementing, and interpreting welfare indicators in macaques, marmosets, and other primate species maintained for biomedical investigation. The content addresses the conceptual basis of welfare assessment, the practical application of behavioral and physiological monitoring tools, and the interpretation of findings within the context of institutional animal care and use programs.

The central question this reference answers is how to distinguish transient adaptive responses from sustained welfare compromise in captive primates. Unlike production species, where welfare assessment protocols have been extensively validated, NHPs present unique challenges: their cognitive complexity, social structure, and species-specific behavioral repertoires demand assessment methods that capture both negative states and positive welfare opportunities. The Guide for the Care and Use of Laboratory Animals establishes the expectation that institutions implement ongoing assessment of animal welfare, yet it does not prescribe specific tools or thresholds. This article addresses that gap by reviewing the scientific foundations and practical options available to the attending veterinarian.

At a Glance

Conceptual Foundations of Welfare Assessment

Welfare is a multidimensional construct encompassing physical health, affective state, and the ability to express species-typical behavior. For NHPs, the cognitive and social dimensions carry particular weight because these animals possess complex neural substrates for emotion, memory, and social cognition. Assessment therefore requires triangulation across domains instead of reliance on any single metric.

The acute phase response provides a physiological foundation for welfare monitoring. Trauma, infection, stress, and inflammation activate a conserved systemic cascade that produces measurable changes in acute phase proteins. As reviewed in the comparative literature on the acute phase response, these proteins have been identified across common animal species and offer applications in diagnosis, prognosis, and laboratory animal welfare assessment. In NHPs, serial measurement of acute phase proteins can detect subclinical inflammation or stress responses before overt clinical signs appear, though interpretation requires species-specific reference intervals and awareness that the response is nonspecific.

The Five Domains Model as an Organizing Framework

The Five Domains model, developed by Mellor and colleagues, organizes welfare assessment into nutrition, environment, health, behavior, and mental state. The model's utility for NHP assessment lies in its explicit separation of physical/functional factors from behavioral expression and its recognition that mental state integrates input from the other four domains. The analysis of welfare insults in neonatal farm animals demonstrates how this framework identifies noxious subjective experiences such as breathlessness, hypothermia, hunger, sickness, and pain, each of which has direct analogues in NHP clinical assessment.

For research primates, the health domain encompasses also disease status but also the consequences of experimental procedures, including surgical recovery, chronic catheterization, and repeated blood sampling. The behavior domain captures both the presence of species-typical activities and the absence of abnormal repetitive behaviors. The mental state domain requires inference from behavioral and physiological evidence, acknowledging that subjective experience cannot be measured directly.

Personality, Coping Style, and Individual Variation

Individual differences in temperament and coping style materially affect welfare indicator interpretation. The review of personality concepts in mammalian farm animals identifies how terminology such as personality, temperament, and coping style are defined and measured, and how these constructs relate to welfare outcomes. Proactive copers typically show active behavioral responses to challenge, while reactive copers show inhibition and withdrawal. These differences mean that a single behavioral threshold for welfare concern will misclassify individuals at the extremes.

In practical terms, the attending veterinarian should establish individual baselines for each NHP under their care. A macaque that normally grooms conspecifics for 30 minutes daily and suddenly ceases social grooming warrants investigation even if all other indicators appear normal. Conversely, an animal with a consistently low activity profile may be expressing its individual phenotype instead of welfare compromise. The Guide for the Care and Use of Laboratory Animals emphasizes that professional judgment, informed by knowledge of the individual animal, is essential to welfare assessment.

Behavioral Indicators

Behavioral assessment forms the primary window into NHP welfare because it is noninvasive, can be performed repeatedly, and captures both negative and positive states. The challenge lies in operationalizing behavioral categories with sufficient precision that different observers achieve acceptable inter-rater reliability.

Species-Typical Behavior as a Baseline

The presence of species-typical behavior indicates that the animal's environment supports normal expression of its behavioral repertoire. For macaques, these behaviors include locomotion, foraging, social grooming, play in juveniles, and appropriate agonistic signaling. For marmosets, scent marking, territorial calling, and cooperative breeding behaviors serve similar functions. Deviation from species-typical patterns, whether in frequency, duration, or context, constitutes the first signal of potential welfare compromise.

Abnormal Repetitive Behaviors

Stereotypies, defined as repetitive, invariant behavioral sequences with no apparent purpose, are among the most visible indicators of compromised welfare in captive NHPs. Common forms include pacing, somersaulting, rocking, and self-directed oral behaviors. The presence of stereotypy indicates that the animal's current environment or history has failed to support normal behavioral development or expression. However, stereotypies can persist after the initiating stressor has been removed, complicating their use as real-time welfare indicators.

Social Behavior and Affiliation

Social withdrawal, increased aggression, or disrupted grooming relationships can signal welfare compromise. Conversely, the presence of allogrooming, play, and tolerant proximity indicates positive social welfare. Assessment of social behavior requires knowledge of the group's dominance hierarchy and individual relationships, since behavior that appears abnormal in one context may be appropriate in another.

Physiological Indicators

Physiological measures provide objective, quantifiable data that complement behavioral observation. Their principal limitation is that many respond to any stressor, making it difficult to attribute changes to welfare compromise versus experimental manipulation or routine handling.

Acute Phase Proteins

The acute phase response offers a window into systemic health status that is more sensitive than overt clinical signs. C-reactive protein, haptoglobin, and fibrinogen are among the proteins that rise following tissue injury, infection, or inflammation. The review of the acute phase response across animal species notes that these proteins have been identified in common laboratory species and hold potential applications for welfare assessment. Serial sampling allows detection of trends that may indicate developing pathology before clinical signs appear.

Hypothalamic-Pituitary-Adrenal Axis Measures

Fecal glucocorticoid metabolites reflect integrated cortisol secretion over hours to days, avoiding the acute elevation caused by venipuncture or restraint. This makes them particularly valuable for assessing chronic stress in NHPs. However, circadian variation, seasonal breeding cycles, and social rank all influence glucocorticoid output, requiring careful experimental design and interpretation. The Guide for the Care and Use of Laboratory Animals identifies physiological measures as one component of a comprehensive assessment program but cautions that no single measure is sufficient.

Body Condition and Physical Examination

Regular physical examination, including body condition scoring, dental assessment, and evaluation of coat condition, provides baseline health data that contextualize other welfare indicators. Weight trends, in particular, offer a sensitive index of overall health that integrates nutritional status, disease burden, and metabolic demand.

Structured Welfare Assessment Protocols

A structured protocol converts the multidimensional indicators described in earlier sections into repeatable, scored observations. The protocol must specify the observation window, the environmental context, the scoring scale, and the threshold that triggers intervention. Without these fixed elements, assessments drift between observers and across time points, and early welfare decline goes undetected.

The observation window should be standardized for each facility. A 10 minute focal observation during the active phase of the light cycle captures foraging, locomotion, and social interaction. A second 5 minute observation after a predictable husbandry event, such as feeding or cage change, reveals anticipatory behavior and response to environmental change. Observations should occur at the same time of day relative to the light cycle, because macaque activity budgets shift markedly between morning and afternoon.

Scoring scales should be anchored to operationally defined behaviors, not to subjective impressions. A 0 to 3 scale for abnormal repetitive behavior, where 0 indicates absence, 1 indicates less than 5 episodes per 10 minutes, 2 indicates 5 to 10 episodes, and 3 indicates continuous or self-injurious performance, allows reliable scoring across trained observers. The same anchoring principle applies to posture, piloerection, and facial expression.

Inter-observer reliability must be established before the protocol enters routine use. Two observers score the same animals independently across at least 10 sessions, and agreement is calculated using weighted kappa. A kappa below 0.6 indicates that the behavioral definitions require revision or that additional training is needed. The Guide for the Care and Use of Laboratory Animals emphasizes that personnel conducting assessments must demonstrate competence in the specific techniques used.

Scoring Sheet Design and Interpretation

The scoring sheet should integrate behavioral, physical, and physiological domains into a single document that generates a composite welfare score. Each domain contributes a weighted subscore, and the weights reflect the domain's predictive validity for the species and the experimental context. Behavioral domains typically carry the highest weight in social species such as macaques because behavior integrates affective state, physical health, and environmental fit.

The composite score serves as a triage tool, not a diagnosis. A score below the intervention threshold prompts increased monitoring frequency. A score at the intervention threshold triggers a veterinary examination and a review of the housing and experimental conditions. A score above the threshold requires immediate action, which may include removal from the study, analgesic intervention, or environmental enrichment modification.

Interpretation guidelines must account for the animal's baseline. An animal with a low baseline activity level due to age or personality may score within normal limits on absolute scales while still experiencing welfare decline relative to its own baseline. The personality research in mammalian farm animals demonstrates that individual differences in coping style and temperament systematically affect behavioral responses to challenge, and the same principle applies to non-human primates. The protocol should therefore include a baseline assessment for each animal before experimental manipulation, and subsequent scores should be interpreted against that individual baseline as well as against species norms.

Decision Points and Escalation Criteria

The protocol must define explicit decision points where the assessment outcome changes the animal's management. These decision points should be written into the animal care and use protocol and approved by the institutional oversight body before the study begins.

The first decision point occurs at the initial post-acclimation assessment. An animal that fails to establish species-typical behavior patterns within the acclimation period, such as persistent huddling, inappetence, or self-directed biting, should not enter the experimental phase until the underlying cause is identified and addressed. The NC3Rs guidance on refinement supports this staged approach, where welfare criteria gate progression through the study timeline.

The second decision point occurs at each scheduled assessment during the experimental phase. A composite score that declines by a defined margin from the individual baseline, or that crosses the absolute intervention threshold, triggers the escalation pathway. The escalation pathway specifies who is notified, what immediate actions are taken, and how quickly the veterinary team must respond.

The third decision point occurs when a physiological indicator, such as a sustained elevation in acute phase protein concentration, accompanies a behavioral decline. The acute phase response review notes that acute phase proteins serve as nonspecific indicators of inflammation, infection, and stress, and their combination with behavioral data strengthens the case for intervention. A single abnormal physiological value should prompt confirmation testing, not immediate action, because sampling stress itself can elevate some parameters.

Equipment and Consumable Selection

The equipment required for welfare assessment depends on the housing system and the physiological parameters being measured. Behavioral observation requires only a timing device and a standardized scoring sheet, but video recording with remote viewing capability reduces observer disturbance and allows retrospective scoring. Infrared cameras support dark-phase observation, which is relevant for nocturnal species.

Physiological sampling requires species-appropriate restraint and collection equipment. Blood collection from macaques typically requires sedation or trained voluntary sampling, and the choice between these approaches affects the validity of the measured parameters. Voluntary sampling from trained animals avoids the stress of capture and chemical restraint, but requires substantial training investment. The MSD Veterinary Manual provides species-specific guidance on blood collection sites and sample handling that applies to non-human primates.

Sample handling affects assay validity. Acute phase protein measurements require serum or plasma separated promptly and stored according to the assay manufacturer's specifications. Fecal glucocorticoid metabolite analysis requires collection of fresh samples, typically within 15 minutes of defecation, and storage at appropriate temperatures to prevent bacterial degradation of the metabolites.

Documentation and Longitudinal Tracking

Documentation serves two functions: it supports individual animal care decisions and it provides the data needed for facility-level welfare monitoring. Each animal should have a welfare record that includes the baseline assessment, all scheduled assessment scores, any unscheduled assessments triggered by clinical signs, and the actions taken in response to threshold crossings.

Longitudinal tracking across the facility identifies patterns that individual assessments miss. A room-level increase in abnormal repetitive behavior scores following a change in husbandry personnel, for example, points to a management issue instead of an individual animal problem. The WOAH terrestrial animal health standards describe surveillance principles that apply to welfare monitoring as well as disease surveillance, including the value of trend analysis over single time point measurements.

The documentation format should support both immediate clinical use and retrospective analysis. Electronic records with structured fields for each scored behavior allow automated calculation of composite scores and trend detection. Paper records are acceptable for small facilities but require manual transcription for analysis, which introduces error risk.

The welfare record should follow the animal through transfers between rooms, studies, or facilities. A macaque that has a documented history of abnormal repetitive behavior during social housing may require a different housing strategy in a subsequent study, and that information must be available to the next veterinary team. The AVMA practice resources emphasize continuity of care as a professional obligation, and this applies with particular force to animals whose welfare status changes across experimental contexts.

Protocol Validation and Refinement

A welfare assessment protocol requires validation before it can be trusted for decision-making. Validation includes demonstrating that the protocol detects known welfare challenges, that scores correlate with independent physiological measures, and that different observers produce consistent results. The review of pain assessment in pigs illustrates the general principle that assessment tools must be validated against known painful or stressful conditions before they are used to make clinical decisions, and the same logic applies to non-human primate welfare assessment.

Validation studies should use a positive control, such as a known stressful event like social separation or veterinary examination, and demonstrate that the protocol detects the expected welfare decline. The protocol should also be tested against a negative control, such as stable social housing with enrichment, to confirm that scores remain within the normal range when welfare is not compromised.

Refinement is an ongoing process. The protocol should be reviewed at defined intervals, typically annually, and after any significant change in housing, husbandry, or experimental procedures. Observer feedback should be incorporated into the review, because the people performing the assessments often identify ambiguous definitions or impractical procedures before they appear in the data. The National Research Council guide describes performance-based standards that allow facilities to adapt their programs to their specific conditions while maintaining defined welfare outcomes.

Recognized Complications and Early Detection

Welfare assessment in non-human primates fails through several recognizable mechanisms. The most common is observer drift, where scoring becomes progressively lenient or inconsistent over weeks of repeated assessment. Detection requires periodic inter-observer reliability checks, ideally with video-recorded sessions scored independently by two trained assessors. A second failure mode is the single-timepoint snapshot, which misses circadian variation in behavior and physiology. HPA axis measures, for example, vary markedly across the day, and a morning sample cannot represent the animal's overall state. Structured protocols that specify time windows for observation, as recommended in the National Research Council guide for laboratory animal care, reduce this risk.

A third complication is the masking of clinical deterioration by compensatory behavior. Non-human primates, particularly macaques, suppress overt signs of pain or illness when observed by humans. Weight loss, reduced grooming, and altered posture may be present for days before overt behavioral change appears. Regular body condition scoring and weighing, performed at intervals short enough to detect trends, are the most reliable early indicators. A fourth failure mode is the misattribution of abnormal behavior to poor welfare when it reflects a learned response or a species-typical pattern. Stereotypic pacing in a singly housed animal warrants concern, but the same behavior in a group-housed animal during a brief disturbance may be transient and adaptive. Discriminating between these requires repeated observation and knowledge of the individual's baseline.

Common Errors and Corrective Action

Less experienced assessors frequently over-rely on a single indicator, most often the presence or absence of abnormal repetitive behavior. This error is compounded when the assessor does not record the frequency, duration, and context of the behavior. A macaque that grooms itself excessively during a 10-minute observation may be expressing a normal behavior at an unusual rate, or may be showing early signs of dermatological disease. The corrective action is to use a structured scoring sheet that forces the assessor to record multiple domains, including posture, locomotion, social interaction, and physical condition, instead of a single behavioral category.

Another common error is the failure to distinguish acute from chronic welfare compromise. Acute pain and distress produce different behavioral and physiological signatures than chronic states, and the two require different assessment tools and different interventions. The review of pain assessment approaches in pigs makes this distinction explicit for that species, and the same principle applies to non-human primates. Acute distress is often detected through vocalisation, grimacing, and altered respiratory rate, while chronic compromise appears as weight loss, poor coat condition, and reduced social engagement. A third error is the collection of physiological samples without behavioral context. A single cortisol measurement obtained after capture and restraint reflects the handling procedure more than the animal's baseline state. The corrective action is to pair physiological sampling with behavioral observation and to interpret results in light of the sampling method.

Limitations of the Current Evidence

The evidence base for welfare assessment in non-human primates is uneven. Behavioral indicators are better characterized than physiological ones, and most validation work has been conducted in macaques, with less information available for marmosets, squirrel monkeys, and other species. The acute phase response literature identifies acute phase proteins as promising markers of inflammation and stress in animals, but species-specific reference ranges and the time course of the response are not fully established for many non-human primate species. Expert opinion differs on the relative weight that should be given to behavioral versus physiological measures, and on whether positive welfare indicators, such as play and social affiliation, should be scored alongside negative indicators. The personality research in farm animals demonstrates that individual variation in coping style can confound group-level welfare assessments, and this finding is likely to apply to non-human primates, but the practical implications for assessment protocols have not been fully worked out.

Referral, Consultation, and Reporting

Veterinarians should seek specialist consultation when an animal's welfare score deteriorates despite standard interventions, when the cause of a behavioral or physical abnormality is unclear, or when a novel enrichment or housing strategy is being considered. Laboratory animal medicine specialists, veterinary behaviorists, and clinical pathologists each bring relevant expertise. Laboratory involvement is indicated when physiological monitoring is required, particularly for serial acute phase protein measurement or HPA axis assessment, because sample handling and assay validation materially affect the results. Regulatory reporting is warranted when welfare compromise results from a protocol violation, an unexpected adverse event, or a housing or husbandry failure that affects multiple animals. Institutional animal care and use committees and the attending veterinarian have overlapping responsibilities in these circumstances, and the NC3Rs guidance on refinement provides a framework for deciding when a protocol should be modified or suspended. The threshold for reporting should be lower than the threshold for intervention, because early reporting allows corrective action before welfare compromise becomes severe.

Frequently Asked Questions

How Can Welfare Be Assessed Reliably When Staff Time and Budget Are Limited?

Prioritize a two-tier system. Tier one uses observations already embedded in daily husbandry, such as appetite, fecal output, grooming quality, and response to familiar staff, recorded on a simple scoring sheet. Tier two applies more detailed behavioral or physiological measures only when tier one flags concern or after procedures. This staged approach concentrates resources where they matter most. The Guide for the Care and Use of Laboratory Animals describes institutional expectations for veterinary care and oversight that can be met without elaborate equipment. Routine physical examination and body condition scoring remain the most cost-effective baseline tools.

What Should Be Done When Recommended Assessment Equipment Is Unavailable?

Use structured observation protocols in place of automated systems. Video recording with later review by a second observer can substitute for real-time scoring and improves inter-observer reliability. For physiological measures, consider sample types that require minimal equipment, such as fecal glucocorticoid metabolites instead of serum cortisol, though collection timing and storage must be standardized. The NC3Rs practical guidance offers refinement strategies that often reduce equipment demands. Document any substitution in the study records and state clearly that the alternative method has known limitations. Do not abandon assessment entirely because ideal tools are absent.

How Do Welfare Indicators Differ Between Macaques and Other NHP Species?

Macaques show prominent abnormal repetitive behaviors such as pacing and self-directed biting, making these useful sentinels. New World primates, including marmosets and squirrel monkeys, display species-specific signs such as piloerection, scent-marking changes, and tremor that are less salient in macaques. Social structure matters: pair-bonded species show different affiliation patterns than multi-male, multi-female groups. The MSD Veterinary Manual provides species-specific clinical and behavioral reference points. Always establish species-typical baselines before interpreting deviations, and consult colony records for historical norms within your own facility.

What Records Must Be Kept for Welfare Assessments to Be Defensible?

Record the date, time, observer identity, animal identification, housing location, and the specific indicators scored. Include the raw data, also the summary score, so that trends can be re-examined later. Note any interventions triggered by the assessment and their outcomes. The AVMA practice resources describe documentation standards that support professional accountability. Keep records in a format that allows longitudinal tracking across the animal's lifetime, also within a single study. If a score triggers escalation, record the clinical reasoning and the personnel consulted.

How Should a Welfare Concern Be Raised With a Principal Investigator or Supervisor?

Present the objective data first, including the scoring sheet, trend over time, and comparison to baseline. State the specific welfare concern and the proposed action, such as analgesia, housing change, or study endpoint. Reference the institutional animal care and use protocol and the WOAH terrestrial animal health standards where relevant. Frame the discussion around study integrity, since unmanaged welfare compromise can confound data. If the concern is urgent and the supervisor is unavailable, follow the institutional chain of command and document the attempt to escalate.

Can Personality Assessment Be Used Practically in Routine Welfare Monitoring?

Yes, but with caution. Stable individual differences in response to novelty, handling, and social challenge can identify animals at higher risk of poor welfare outcomes. The review of personality concepts and measures in farm animals illustrates how temperament traits correlate with welfare-relevant responses. In practice, assign each animal a simple coping style category after repeated observations, then use that category to interpret subsequent behavioral signs. A timid animal showing reduced exploration may be distressed, while the same behavior in a bold animal is more alarming. Personality data should supplement, not replace, direct welfare indicators.

Related Clinical & Scientific Guides

• Refining IACUC Protocols to Minimize Animal Pain and Distress
• Health Monitoring Programs for Laboratory Animal Facilities
• Anesthetic Risk Assessment in Laboratory Animals: Preoperative Evaluation

References and Further Reading

• Assessing learning and memory in pigs.. 2011.
• Personality Research in Mammalian Farm Animals: Concepts, Measures, and Relationship to Welfare.. 2018.
• A Review of Pain Assessment in Pigs.. 2016.
• Acute phase response in animals: a review.. 2009.
• Public health risk of antimicrobial resistance transfer from companion animals.. 2017.
• Animal welfare implications of neonatal mortality and morbidity in farm animals.. 2004.
• Guide for the Care and Use of Laboratory Animals, 8th Edition. National Academies Press, 2011.
• NC3Rs Resources on Replacement, Reduction and Refinement. NC3Rs.
• MSD Veterinary Manual, Professional Edition. MSD Veterinary Manual.

Related Articles

• Welfare Assessment Tools for Laboratory Rodents
• Training and Competency Assessment for Animal Research Personnel
• Welfare Assessment of Laboratory Fish: Beyond Zebrafish
• Zebrafish Welfare in Research: Housing and Environmental Needs
• Animal Model Selection for Neurological Research

This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.