The New Hampshire Board of Pharmacy is considering proposed rules (Part Ph 522) that would establish detailed standards for the use of Artificial Intelligence (AI) and other technology in pharmacy practice. This marks an early and significant effort by a state pharmacy board to create a specific regulatory framework, addressing how AI can support, but not replace, professional judgment in areas like prescription processing, patient communication, drug utilization review, and dispensing.
AI May Assist Pharmacy Practice, But Cannot Replace Professional Judgment The core tenet of the proposed New Hampshire rules is that technology and AI should only serve as support tools for pharmacy practice, never replacing the essential professional judgment, responsibility, or accountability of licensed pharmacy personnel. Pharmacists would retain full responsibility for critical decisions such as prospective drug utilization review, patient counseling, clinical assessments, and final prescription verification. The rules specifically prohibit AI from automatically verifying, approving, dispensing, or suppressing clinical information in a manner that bypasses legally mandated pharmacist review or judgment. This establishes a clear boundary between AI as an aid and AI as a decision-maker in a professional context. Pharmacies Would Have Significant Pre-Implementation Obligations Before integrating any technology-assisted workflow that significantly impacts prescription processing, dispensing, verification, patient communications, drug utilization review, or clinical decision support, pharmacies would be subject to extensive pre-implementation requirements. These include thoroughly evaluating the technology's intended use, identifying authorized users, validating its intended function, and establishing comprehensive written policies and procedures. Furthermore, robust training for personnel on proper usage, limitations, and escalation protocols would be mandatory. For AI and clinical decision support tools, pharmacies must also document processes for accepting, rejecting, or overriding system recommendations and establish procedures for monitoring performance, errors, bias, and unintended consequences, signaling a proactive approach to risk management. Auditability Could Become a Key Compliance Requirement A significant emphasis of the proposed regulations is on record-keeping and auditability. Technology-assisted systems would be required to maintain detailed records for at least four years, capturing information on the patient, prescription, date, time, and specific actions taken. Crucially, these records must identify individuals who performed, reviewed, approved, modified, or overrode activities, along with any required pharmacist reviews or interventions. For AI, the system must record when recommendations, alerts, or other outputs are accepted, rejected, modified, or overridden, and attribute the decision to the responsible individual. This requirement implies that AI tools must be designed with robust audit trails to ensure compliance, potentially influencing vendor selection for pharmacies. Vendor Contracts May Need Greater Scrutiny The proposal explicitly states that outsourcing technology does not absolve the pharmacy permit holder of regulatory responsibility. Pharmacies utilizing third-party technology for their operations must ensure that vendor solutions comply with all regulatory obligations, particularly concerning confidentiality, auditability, record retention, security, downtime procedures, and regulatory inspection. This means pharmacy contracts with AI developers, pharmacy management system vendors, and other technology providers may need to include specific clauses granting contractual rights related to access to audit logs, data retention, system validation, cybersecurity measures, incident notification, and cooperation with regulators, shifting the onus onto pharmacies to secure these provisions. AI Governance May Become Part of Pharmacy Quality Assurance The proposed rules would integrate AI governance into the ongoing quality assurance and compliance infrastructure of pharmacies. Technology-assisted workflows would require periodic reviews – before implementation, following material changes, after certain errors or security incidents, and at least annually. Pharmacies would be mandated to document these reviews and any subsequent corrective actions. This ensures that the use of AI is not a one-time implementation event but an continuous process of monitoring and improvement, requiring pharmacies to maintain vigilance over the performance and impact of their AI systems as part of their routine operational oversight. The Proposal Reaches Beyond Generative AI The proposed definition of artificial intelligence in Ph 522 is notably broad, encompassing machine-based systems or software tools that use algorithms, machine learning, natural language processing, predictive analytics, or similar functionalities to generate various outputs. This extensive definition suggests that the regulations would apply to a wide range of existing and future technologies beyond just generative AI products like conversational assistants. It could potentially impact tools used for drug utilization review, adherence prediction, prescription prioritization, clinical interventions, patient outreach, fraud detection, and other forms of algorithmic decision support, requiring pharmacies to assess a broad spectrum of their current technology landscape for compliance. Practical Considerations for Pharmacy Operators Pharmacy operators, especially those with national reach, should view the New Hampshire proposal as a potential blueprint for future state-level AI regulations, given that boards often learn from other jurisdictions. It is advisable for pharmacies to begin inventorying all AI and clinical decision support tools currently in use, paying close attention to those that influence dispensing, drug utilization review, patient prioritization, or access to medications. Furthermore, organizations should evaluate their existing governance structures to ensure they address key themes in the proposal, such as human oversight, validation, bias and error monitoring, cybersecurity, vendor oversight, training, downtime procedures, and periodic quality review, to prepare for evolving regulatory expectations. What Next? The New Hampshire Board of Pharmacy continues its work on the comprehensive Ph 500 standards-of-practice rulemaking, with Part Ph 522 poised to establish a detailed framework for technology-assisted pharmacy practice, including a dedicated section for AI and clinical decision support. Stakeholders, including pharmacies and technology vendors, are strongly advised to monitor the rulemaking process closely. This proposal is significant as it directly incorporates AI regulation into the traditional framework of pharmacy oversight, emphasizing professional judgment, permit-holder accountability, robust record-keeping, and continuous quality assurance, setting a precedent for how AI will be governed in healthcare.
AI May Assist Pharmacy Practice, But Cannot Replace Professional Judgment
The core tenet of the proposed New Hampshire rules is that technology and AI should only serve as support tools for pharmacy practice, never replacing the essential professional judgment, responsibility, or accountability of licensed pharmacy personnel. Pharmacists would retain full responsibility for critical decisions such as prospective drug utilization review, patient counseling, clinical assessments, and final prescription verification. The rules specifically prohibit AI from automatically verifying, approving, dispensing, or suppressing clinical information in a manner that bypasses legally mandated pharmacist review or judgment. This establishes a clear boundary between AI as an aid and AI as a decision-maker in a professional context.
Pharmacies Would Have Significant Pre-Implementation Obligations
Before integrating any technology-assisted workflow that significantly impacts prescription processing, dispensing, verification, patient communications, drug utilization review, or clinical decision support, pharmacies would be subject to extensive pre-implementation requirements. These include thoroughly evaluating the technology's intended use, identifying authorized users, validating its intended function, and establishing comprehensive written policies and procedures. Furthermore, robust training for personnel on proper usage, limitations, and escalation protocols would be mandatory. For AI and clinical decision support tools, pharmacies must also document processes for accepting, rejecting, or overriding system recommendations and establish procedures for monitoring performance, errors, bias, and unintended consequences, signaling a proactive approach to risk management.
Auditability Could Become a Key Compliance Requirement
A significant emphasis of the proposed regulations is on record-keeping and auditability. Technology-assisted systems would be required to maintain detailed records for at least four years, capturing information on the patient, prescription, date, time, and specific actions taken. Crucially, these records must identify individuals who performed, reviewed, approved, modified, or overrode activities, along with any required pharmacist reviews or interventions. For AI, the system must record when recommendations, alerts, or other outputs are accepted, rejected, modified, or overridden, and attribute the decision to the responsible individual. This requirement implies that AI tools must be designed with robust audit trails to ensure compliance, potentially influencing vendor selection for pharmacies.
Vendor Contracts May Need Greater Scrutiny
The proposal explicitly states that outsourcing technology does not absolve the pharmacy permit holder of regulatory responsibility. Pharmacies utilizing third-party technology for their operations must ensure that vendor solutions comply with all regulatory obligations, particularly concerning confidentiality, auditability, record retention, security, downtime procedures, and regulatory inspection. This means pharmacy contracts with AI developers, pharmacy management system vendors, and other technology providers may need to include specific clauses granting contractual rights related to access to audit logs, data retention, system validation, cybersecurity measures, incident notification, and cooperation with regulators, shifting the onus onto pharmacies to secure these provisions.
AI Governance May Become Part of Pharmacy Quality Assurance
The proposed rules would integrate AI governance into the ongoing quality assurance and compliance infrastructure of pharmacies. Technology-assisted workflows would require periodic reviews – before implementation, following material changes, after certain errors or security incidents, and at least annually. Pharmacies would be mandated to document these reviews and any subsequent corrective actions. This ensures that the use of AI is not a one-time implementation event but an continuous process of monitoring and improvement, requiring pharmacies to maintain vigilance over the performance and impact of their AI systems as part of their routine operational oversight.
The Proposal Reaches Beyond Generative AI
The proposed definition of artificial intelligence in Ph 522 is notably broad, encompassing machine-based systems or software tools that use algorithms, machine learning, natural language processing, predictive analytics, or similar functionalities to generate various outputs. This extensive definition suggests that the regulations would apply to a wide range of existing and future technologies beyond just generative AI products like conversational assistants. It could potentially impact tools used for drug utilization review, adherence prediction, prescription prioritization, clinical interventions, patient outreach, fraud detection, and other forms of algorithmic decision support, requiring pharmacies to assess a broad spectrum of their current technology landscape for compliance.
Practical Considerations for Pharmacy Operators
Pharmacy operators, especially those with national reach, should view the New Hampshire proposal as a potential blueprint for future state-level AI regulations, given that boards often learn from other jurisdictions. It is advisable for pharmacies to begin inventorying all AI and clinical decision support tools currently in use, paying close attention to those that influence dispensing, drug utilization review, patient prioritization, or access to medications. Furthermore, organizations should evaluate their existing governance structures to ensure they address key themes in the proposal, such as human oversight, validation, bias and error monitoring, cybersecurity, vendor oversight, training, downtime procedures, and periodic quality review, to prepare for evolving regulatory expectations.
What Next?
The New Hampshire Board of Pharmacy continues its work on the comprehensive Ph 500 standards-of-practice rulemaking, with Part Ph 522 poised to establish a detailed framework for technology-assisted pharmacy practice, including a dedicated section for AI and clinical decision support. Stakeholders, including pharmacies and technology vendors, are strongly advised to monitor the rulemaking process closely. This proposal is significant as it directly incorporates AI regulation into the traditional framework of pharmacy oversight, emphasizing professional judgment, permit-holder accountability, robust record-keeping, and continuous quality assurance, setting a precedent for how AI will be governed in healthcare.
