What Resource-Use Data Should Be Collected During an Intervention Study?

Which Resources Should Researchers Record When Evaluating an Intervention?

A university evaluates a digital academic support intervention. Researchers collect learning outcomes, participation records, and student satisfaction data. They also record the software subscription fee, assuming that this represents the cost of implementing the program.

However, the intervention requires instructor preparation, technical support, student orientation, additional administrative work, and the use of existing institutional equipment. Some students also incur connectivity expenses. These resources are not necessarily visible in the software invoice.

If the researchers later want to estimate implementation costs or compare the program economically with existing practice, the missing information may limit their conclusions.

What resource-use data should have been collected, and how can researchers identify the relevant information without creating an unnecessarily demanding data collection process?

Collect Resource Quantities That Are Relevant to the Economic Question and Perspective

Identifying, Measuring, and Valuing Resources Used During Intervention Research

What Counts as Resource Use in an Intervention Study?

Resource use refers to the quantities of inputs consumed in developing, introducing, delivering, monitoring, or maintaining an intervention, as well as other resource consequences relevant to the research question.

These inputs may include staff hours, training sessions, equipment use, software licenses, clinical services, travel, facilities, supplies, and participant time.

For example, an educational intervention may require 40 instructor-hours, 12 hours of technical support, and access to an institutional learning platform. These quantities describe resource consumption even before monetary values are assigned.

Resource-use measurement should be distinguished from expenditure reporting. A resource may have economic value even when the organization does not make an additional payment during the study.

Researchers should identify the resources relevant to their economic question rather than equate resource use with recorded invoices.

How Is Resource Use Different From Cost?

Resource use concerns quantities. Cost concerns the monetary value assigned to those quantities using an appropriate valuation method.

For example, recording that a research intervention required 25 hours of instructor preparation provides resource-use information. Multiplying those hours by a justified unit cost produces a monetary estimate.

Separating these stages is useful because unit costs may differ between institutions, countries, or time periods.

Researchers should preserve information about quantities and valuation assumptions so that readers can understand how cost estimates were constructed.

The broader question of when cost and resource use should be measured should be resolved before developing a detailed resource inventory.

How Does the Economic Perspective Determine Which Resources to Collect?

The economic perspective identifies whose resource consequences are relevant to the analysis.

An institutional perspective may include staff time, infrastructure, software, and operating expenditure. A healthcare payer perspective may focus on reimbursed services. A societal perspective may require broader consideration of resource consequences, including relevant participant time and productivity losses.

For example, a university may provide a digital learning platform without charging students directly. Nevertheless, students may incur additional connectivity or equipment expenses.

Those expenses may be excluded from a narrowly defined institutional costing analysis but relevant to a broader evaluation.

Researchers should determine whether participant costs belong alongside institutional costs before selecting measurement instruments.

Which Resource Categories Should Researchers Consider?

The following categories provide a starting point for identifying resources. Their inclusion should depend on the intervention and analytical perspective.

Researchers should avoid including every possible category merely because it appears in a resource-use framework. The purpose is to capture relevant differences between the alternatives being evaluated.

How Should Personnel Time Be Measured?

Personnel time can be a substantial component of intervention resource use, particularly in education, healthcare, and community-based programs.

Researchers should identify relevant staff categories and distinguish activities such as preparation, training, direct delivery, supervision, administration, and technical support.

Time may be measured through activity logs, timesheets, structured observation, scheduling records, or other suitable methods.

For example, an instructor's time spent preparing digital feedback may need to be distinguished from ordinary teaching responsibilities already required under the comparator.

Researchers should avoid assuming that all scheduled staff time is incremental. The economic question may concern additional resources compared with existing practice rather than total staff workload.

Should Unpaid or Existing Staff Time Be Recorded?

Potentially, yes. The absence of additional expenditure does not necessarily mean that no resource was consumed.

For example, an existing instructor may deliver an intervention during hours already covered by an institutional salary. Those hours may nevertheless displace other activities.

Whether the time should be assigned a monetary value depends on the perspective and costing method.

Researchers should distinguish financial expenditure from economic opportunity cost, which reflects the value of resources in their next-best alternative use.

They should also avoid double counting staff time already included in another cost category.

How Should Equipment and Infrastructure Be Recorded?

Equipment and infrastructure should be measured according to their contribution to the intervention and the relevant time horizon.

For example, a digital intervention may use computers already owned by an institution. Recording the full purchase price of every computer as an intervention cost would usually be inappropriate when the equipment is shared across multiple activities.

Researchers may need to consider useful life, intensity of use, shared utilization, and appropriate allocation methods.

One-time capital expenditure should be distinguished from recurring operating costs.

Where equipment is required exclusively for the intervention, the relevant economic treatment should still reflect the perspective and analytical period.

What Software and Digital Resource Data Matter?

Digital interventions may require licensing, hosting, storage, technical maintenance, cybersecurity services, user support, and platform integration.

Researchers should identify whether charges are fixed, per-user, usage-based, or dependent on organizational scale.

For example, a platform may charge a fixed annual institutional fee plus an additional amount for each active user.

Resource records should distinguish setup activities from continuing service requirements.

Digital systems may also generate useful administrative records of intervention exposure, but these records do not automatically capture staff effort or the full cost of operating the system.

Should Training and Implementation Activities Be Included?

Training and implementation support may be necessary to introduce an intervention successfully.

Relevant resources may include trainer preparation, staff attendance, instructional materials, supervision, implementation facilitation, and workflow redesign.

These resources should be distinguished from the direct activities through which participants receive the intervention.

For example, training instructors to use a formative feedback system is an implementation activity, while providing feedback to students is part of intervention delivery.

The distinction matters when estimating the cost of initial adoption separately from the cost of continued operation.

Researchers should consider cost and other implementation requirements when the intervention is intended for routine adoption.

How Should Research-Specific Activities Be Distinguished From Intervention Delivery?

Research-specific activities are procedures undertaken to conduct the study rather than to deliver the intervention in ordinary practice.

They may include additional outcome assessments, research interviews, trial-specific monitoring, and administrative procedures required only for evaluation.

These activities can consume substantial resources but may not belong in an estimate of routine intervention delivery cost.

However, some monitoring or supervision may remain necessary during ordinary implementation. Such resources should not be excluded merely because they were also used during the study.

Researchers should classify each activity according to its actual function and the economic question being answered.

Should Resource Use Be Measured in the Comparison Group?

Yes, when the objective is to estimate comparative costs or economic value.

An intervention's economic consequences cannot be determined adequately from intervention-group resource use alone if the comparator also consumes resources.

For example, a new digital tutoring program may require additional software expenditure but reduce the need for existing face-to-face support sessions.

Measuring only the new program's costs would omit potentially important differences in the resources used by the alternatives.

Researchers should use consistent resource definitions and observation periods across groups where appropriate.

What Are Incremental Resource Requirements?

Incremental resource requirements are the differences in resource use between the intervention and its relevant comparator.

For example, an intervention may require additional instructor preparation but fewer administrative follow-up contacts.

Both differences may be relevant to the economic comparison.

Researchers should avoid assuming that every resource consumed by the intervention represents an additional resource requirement.

When costs are valued, incremental cost can be calculated as the difference between appropriately estimated costs for the alternatives.

How Should Resource-Use Data Be Collected?

Several approaches may be appropriate, depending on the resources and available information.

Combining sources may improve coverage, but researchers should establish procedures for resolving overlap and preventing double counting.

When Should Resource Use Be Measured?

Measurement timing should reflect when relevant resources are consumed.

Initial implementation may involve concentrated training and setup activities, while routine delivery may generate recurring personnel and operating requirements.

Downstream resource use may occur after the intervention has ended.

Researchers should select an appropriate economic time horizon and determine which resources must be observed during each period.

Not every resource needs to be measured at every research visit. For example, software expenditure may be documented through invoices, while participant travel costs may require reporting after relevant appointments.

How Do Recall Periods Affect Resource-Use Measurement?

Recall periods determine how far back participants or providers must remember resource use.

Shorter recall periods may improve reporting of frequent or minor events, but they can increase the number of assessments.

Longer periods may reduce contact burden while increasing the possibility of forgotten activities or inaccurate quantities.

The appropriate interval depends on the resource, its frequency, its salience, and the availability of independent records.

Researchers should evaluate recall requirements during feasibility testing rather than assume one interval is optimal for every resource category.

How Can Resource-Use Questionnaires Be Designed?

Questionnaires should define resource categories clearly and use response formats appropriate to the intended population.

For example, asking participants about the number of transportation trips may be easier than requesting an unstructured estimate of all travel expenses over several months.

Researchers should avoid ambiguous categories that combine different services or activities.

Where possible, instruments should be informed by existing measurement evidence and tested for comprehension, recall, and usability.

Researchers should not assume that a questionnaire validated in one healthcare system or population will necessarily perform similarly in a different country or educational setting.

Should Participant Time and Productivity Losses Be Recorded?

Participant time and productivity losses may be relevant when the selected perspective includes these consequences.

For example, an intervention may require students to attend sessions outside their ordinary class schedules or employees to take time away from paid work.

Researchers should distinguish time spent receiving the intervention from time spent completing research-specific procedures.

They should also distinguish actual lost earnings from the economic valuation of time.

The treatment of productivity losses and time costs depends on the economic framework and requires attention to potential double counting.

How Should Shared Resources Be Allocated?

Shared resources are used for more than one intervention, department, or organizational activity.

For example, a university learning platform may support several courses, while technical support staff serve multiple programs.

Researchers should establish an allocation method that reflects the relevant resource consumption as closely as practicable.

Possible allocation bases include actual usage, staff time, participant numbers, or another justified measure.

Different allocation assumptions may produce different cost estimates. Researchers should document these assumptions and consider sensitivity analyses when they materially influence results.

How Should Resource-Use Data Be Valued?

Once resource quantities have been measured, appropriate unit costs can be assigned.

For example, staff hours may be valued using a justified labor cost, while service episodes may be valued using suitable unit-cost sources.

Researchers should distinguish prices, charges, reimbursements, and economic costs because these may represent different quantities.

The selected valuation method should match the perspective and research question.

Cost reports should specify the currency, price year, unit-cost sources, and relevant allocation or adjustment assumptions.

How Can Researchers Avoid Double Counting?

Double counting occurs when the same resource or economic consequence is included more than once.

For example, a service contract may already include technical support. Adding the full contract price and separately valuing the same support hours could overstate costs.

Similarly, a staff salary allocation may already cover time recorded in another personnel category.

Researchers should map resource categories to their underlying activities and examine whether bundled costs overlap with separately measured inputs.

Where categories overlap, the accounting treatment should be explained and consistently applied.

How Should Missing Resource-Use Data Be Handled?

Resource-use data may be missing because participants do not complete questionnaires, staff fail to maintain logs, or administrative systems do not record certain activities.

Missingness may be related to participant characteristics, intervention intensity, or the burden of data collection.

Researchers should monitor completeness and identify which resource categories are most affected.

Appropriate statistical methods may be used under defensible assumptions, but they cannot guarantee recovery of unobserved information.

Where conclusions depend on missing-data assumptions, sensitivity analyses should be considered.

How Can Researchers Assess Data Quality?

Data quality assessment should examine completeness, consistency, plausibility, and alignment with the intended resource definitions.

For example, unusually high reported staff hours may reflect genuine intensive delivery, duplicate entries, or misunderstanding of the reporting period.

Researchers should establish procedures for checking records and resolving discrepancies without automatically deleting unusual observations.

Where multiple sources describe the same activity, their definitions and observation periods should be compared.

Data validation procedures should be documented prospectively when feasible.

How Can Economic Data Collection Be Made Less Burdensome?

Resource-use measurement should be proportionate to the economic question.

Researchers may prioritize resources expected to contribute substantially to costs or differ meaningfully between alternatives.

Routine records can reduce direct reporting demands when they contain suitable information.

Questionnaires may be reserved for resources that cannot be captured reliably through other sources.

Researchers should consider the additional burden of economic data collection and avoid repeated requests for information already available through approved records.

What Privacy and Governance Requirements Apply?

Resource-use information may include sensitive details about healthcare utilization, employment, household expenses, or institutional operations.

Researchers should collect only information necessary for the stated objectives and apply appropriate confidentiality, access, retention, and security arrangements.

Administrative record linkage may require institutional authorization and applicable ethics or data protection review.

Requirements differ by jurisdiction and the nature of the information collected.

Economic usefulness does not remove the obligation to justify personal data collection.

How Should Resource-Use Measurement Be Reported?

Reports should identify the perspective, resource categories, measurement instruments, data sources, observation periods, and valuation methods.

Researchers should distinguish resource quantities from costs and explain how shared resources and research-specific activities were treated.

Missing data and uncertainty should be described transparently.

For health economic evaluations, CHEERS 2022 provides relevant reporting guidance, while ISPOR good research practice publications address economic evaluations conducted alongside trials.

These resources support reporting and methodological planning but do not replace study-specific decisions about which inputs matter.

Developing a Resource-Use Inventory for a Digital Learning Intervention

The ₱710 average represents the cost of the specified resources under the hypothetical assumptions. It does not necessarily represent the incremental cost relative to existing practice or the cost of delivering the program at a different scale.

To conduct a comparative economic evaluation, the team would also need relevant comparator resource use and appropriately measured outcomes.

Common Misconceptions About Intervention Resource-Use Data

How to Build a Defensible Resource-Use Measurement Plan

Start by defining the economic question and identifying the resources consumed by the intervention and its comparator. Separate quantities from valuations, select appropriate sources, and establish consistent measurement periods.

Prioritize resources that could materially influence economic conclusions while maintaining sufficient coverage of relevant costs.

Resource-use measurement should be designed before important activities occur whenever possible. A well-defined inventory is often more useful than an extensive collection of poorly classified expenditure records.

Check Whether Your Intervention Resource-Use Data Are Complete and Appropriate

Frequently Asked Questions About Intervention Resource-Use Measurement

Collect the Resources Needed for a Credible Economic Comparison

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.