Deze pagina is automatisch vertaald en de nauwkeurigheid van de vertaling kan niet worden gegarandeerd. Raadpleeg de Engelse versie voor een brontekst.

AI Voice Calls Versus Text and Email Reminders for Questionnaire Response After Shoulder Arthroplasty (VOICE-REACH)

Proactive Conversational Artificial Intelligence Calls Versus Digital Reminders to Collect Patient-Reported Outcomes After Shoulder Arthroplasty: A Randomized Study Within a Multicenter Cohort

After shoulder replacement surgery, surgeons send questionnaires to their patients to learn how they are doing (pain, shoulder function, satisfaction). Many patients do not answer. This study tests the best way to obtain these answers.

The study is nested in a multicenter study of pain after shoulder arthroplasty conducted in four French centers (host study RetroPAinRSACAll, supported by Ramsay Santé). The clinical results of the host study are not the subject of this registration: this study evaluates only the method used to collect the information.

All patients first receive an email invitation (when an email address is available) and, the next day, a text message (SMS) with a personal link to the questionnaire and a phone number to answer at any time by talking with an automated voice assistant. Patients who have not answered two days after the first contact are randomly assigned to one of two follow-up strategies: (1) further digital reminders (SMS on day 3 and day 5, email on day 7), or (2) up to three phone calls from a conversational artificial intelligence (AI) voice agent on days 3, 5 and 7, which asks the same questions orally. In both groups, patients can answer at any time online or by calling the voice assistant, and contacts stop as soon as the questionnaire is completed or the patient refuses.

The main outcome is the proportion of patients who complete the questionnaire within 14 days of the first contact.

Studie Overzicht

Gedetailleerde beschrijving

Design. VOICE-REACH is a pragmatic, multicenter, individually randomized, two-arm, open-label superiority study within a trial (SWAT). It is embedded in the host study RetroPAinRSACAll (persistent pain after shoulder arthroplasty, four French shoulder surgery centers, favorable opinion of the Ramsay Santé scientific committee, reference COS-RGDS-2026-07-008). All patients of the host cohort are contacted through the DeepCall platform on behalf of their surgeon to complete the host study questionnaire (operation details, warning signs, pain screening and characterization, rehabilitation and function including the Subjective Shoulder Value, opinion on the operation and on the contact). The comparison concerns the follow-up strategy only: the questionnaire is the same in both arms and the same on the web and by voice.

Common phase (all participants). Day 0: email invitation with a personal questionnaire link, sent when an email address is available. Day 1: SMS with the same link, the phone number of the AI voice agent (available 24 hours a day for inbound calls) and an opt-out option (reply NO). Patients are contacted in successive waves of up to 450 patients, one day apart, to respect operator sending limits; all study days are counted from each participant's own day 0.

Randomization. On day 2, participants who have not completed the questionnaire and have not refused are randomized 1:1 by the platform, using a computer-generated permuted block sequence (blocks of 4) stratified by center. Participants who completed the questionnaire before day 2 are not randomized and are described separately.

Arm A, digital reminders: SMS reminders on day 3 and day 5 and an email reminder on day 7 (when an email address is available).

Arm B, proactive AI calls: outbound calls from the AI voice agent on day 3, day 5 and day 7 (up to three attempts). The agent calls on behalf of the surgeon's team and asks the questionnaire orally; answers are extracted into the same structured fields as the web form.

In both arms, the web link and the inbound voice line remain available throughout, and all scheduled contacts stop as soon as the questionnaire is completed, the patient refuses, or the patient is identified as deceased. Warning signs reported by patients are flagged to the surgical team for review; the AI agent does not give medical advice.

Analysis. The primary analysis is intention to treat among all randomized participants: proportion with a completed questionnaire by day 14, risk difference (Arm B minus Arm A) with 95% confidence interval, adjusted for center and email availability. Participants without a completed questionnaire by day 14 are counted as non-completers; no answer is imputed. Secondary analyses describe the yield of the common phase, time to completion, completion channel, partial responses, refusals, number of contacts and cost per completed questionnaire, acceptability of the contact mode among responders, representativeness of responders, and the accuracy of voice-captured answers.

Registration timing. The common pre-randomization outreach started on 9 November 2026. This record was created on 10 November 2026, before the first randomization (day 2 of the first wave, 11 November 2026).

Studietype

Ingrijpend

Inschrijving (Geschat)

846

Fase

  • Niet toepasbaar

Contacten en locaties

In dit gedeelte vindt u de contactgegevens van degenen die het onderzoek uitvoeren en informatie over waar dit onderzoek wordt uitgevoerd.

Studiecontact

Studie Locaties

      • Caen, Frankrijk, 14000
        • Werving
        • Hôpital Privé Saint-Martin
        • Hoofdonderzoeker:
          • Franck Dordain, MD
      • Lyon, Frankrijk, 69008
        • Werving
        • Centre Orthopédique Santy, Hôpital Privé Jean Mermoz
        • Hoofdonderzoeker:
          • Arnaud Godenèche, MD
      • Paris, Frankrijk, 75019
        • Werving
        • Clinique des Maussins-Nollet
        • Hoofdonderzoeker:
          • Geoffroy Nourissat, MD, PhD
      • Strasbourg, Frankrijk, 67000
        • Werving
        • Clinique de l'Orangerie
        • Hoofdonderzoeker:
          • Maxime Antoni, MD

Deelname Criteria

Onderzoekers zoeken naar mensen die aan een bepaalde beschrijving voldoen, de zogenaamde geschiktheidscriteria. Enkele voorbeelden van deze criteria zijn iemands algemene gezondheidstoestand of eerdere behandelingen.

Geschiktheidscriteria

Leeftijden die in aanmerking komen voor studie

18 jaar en ouder (Volwassen, Oudere volwassene)

Accepteert gezonde vrijwilligers

Nee

Geslachten die in aanmerking komen voor studie

Allemaal

Beschrijving

Inclusion Criteria:

  • Age 18 years or older
  • Shoulder arthroplasty (reverse, anatomic, humeral head replacement or revision arthroplasty) performed at one of the four participating centers and listed in the host study cohort
  • At least one telephone number available (an email address is not required)

Additional criteria for randomization (day 2):

* Questionnaire not completed and no refusal at the day-2 cutoff

Exclusion Criteria:

  • Death identified before the first contact (cross-check with the French national death registry, INSEE)
  • Duplicate record
  • Refusal of contact
  • Questionnaire already completed before day 2 (excluded from randomization only, described separately)

Studie plan

Dit gedeelte bevat details van het studieplan, inclusief hoe de studie is opgezet en wat de studie meet.

Hoe is de studie opgezet?

Ontwerpdetails

  • Primair doel: Onderzoek naar gezondheidsdiensten
  • Toewijzing: Gerandomiseerd
  • Interventioneel model: Parallelle opdracht
  • Masker: Geen (open label)

Wapens en interventies

Deelnemersgroep / Arm
Interventie / Behandeling
Actieve vergelijker: Digital Reminders

Common phase: email invitation on day 0 (if an email address is available) and SMS invitation on day 1 with the questionnaire link and the AI voice agent phone number. After randomization on day 2: SMS reminder on day 3, SMS reminder on day 5 and email reminder on day 7 (if an email address is available). Participants can answer at any time by web link or by calling the AI voice agent. Contacts stop at completion or refusal.

Standardized reminder messages sent by the DeepCall platform on behalf of the surgeon: SMS on day 3 and day 5 and email on day 7 (if an email address is available), each with the personal questionnaire link, the AI voice agent phone number and an opt-out option.

Experimenteel: Proactive AI Voice Calls

Same common phase (email on day 0 if available, SMS on day 1). After randomization on day 2: up to three outbound calls from the conversational AI voice agent on day 3, day 5 and day 7, during which the same questionnaire is asked orally. Participants can also answer at any time by web link or by calling the AI voice agent. Contacts stop at completion or refusal.

Outbound telephone calls placed by a conversational AI voice agent (DeepCall platform) on day 3, day 5 and day 7, up to three attempts. The agent calls on behalf of the surgeon's team, conducts the same questionnaire orally and records the answers in the same structured fields as the web form.

Andere namen:
  • DeepCall

Wat meet het onderzoek?

Primaire uitkomstmaten

Uitkomstmaat
Maatregel Beschrijving
Tijdsspanne
Questionnaire Completion by Day 14 Among Randomized Participants
Tijdsspanne: From day 0 (first contact) to day 14

Proportion of randomized participants with a completed questionnaire within 14 days of their first contact (day 0), whatever the channel used (web link, inbound call or outbound AI call). A questionnaire is complete when it was submitted on the web or when all required items were answered during a voice interview. Participants without a completed questionnaire at the day-14 cutoff are counted as non-completers. Compared between arms (proactive AI voice calls minus digital reminders).

From day 0 (first contact) to day 14

Secundaire uitkomstmaten

Uitkomstmaat
Maatregel Beschrijving
Tijdsspanne
Questionnaire Completion by Day 7 Among Randomized Participants
Tijdsspanne: From day 0 to day 7

Proportion of randomized participants with a completed questionnaire by day 7, whatever the channel used.

From day 0 to day 7
Time to Questionnaire Completion
Tijdsspanne: From day 2 to day 14

Time in hours from randomization (day 2) to questionnaire completion among randomized participants; participants without a completed questionnaire are censored at day 14.

From day 2 to day 14
Questionnaire Completion Before Randomization
Tijdsspanne: From day 0 to day 2

Proportion of all enrolled participants who completed the questionnaire after the common email (day 0) and SMS (day 1) invitations, before the day-2 randomization.

From day 0 to day 2
Questionnaire Completion by Day 14 in All Enrolled Participants
Tijdsspanne: From day 0 to day 14

Proportion of all enrolled participants, randomized or not, with a completed questionnaire by day 14.

From day 0 to day 14
Partial Questionnaire Response
Tijdsspanne: From day 0 to day 14

Proportion of randomized participants who answered at least one question without completing the questionnaire by day 14.

From day 0 to day 14
Completion Channel
Tijdsspanne: From day 0 to day 14

Distribution of completed questionnaires by channel (web link, inbound call to the voice agent, outbound AI call), by arm.

From day 0 to day 14
Refusal of Contact
Tijdsspanne: From day 2 to day 14

Proportion of randomized participants who explicitly refused further contact (SMS reply, email refusal link or refusal during a call).

From day 2 to day 14
Number of Contacts per Completed Questionnaire
Tijdsspanne: From day 0 to day 14

Number of emails, SMS and outbound calls sent or attempted, divided by the number of completed questionnaires, by arm.

From day 0 to day 14
Direct Cost per Completed Questionnaire
Tijdsspanne: From day 0 to day 14

Direct costs of messaging, telephony and AI processing (euros) divided by the number of completed questionnaires, by arm, and incremental cost per additional completed questionnaire.

From day 0 to day 14
Satisfaction With the Mode of Contact
Tijdsspanne: From day 0 to day 14

Among participants who completed the questionnaire: satisfaction with the way they were contacted, rated on a 5-level scale from very satisfied to very dissatisfied.

From day 0 to day 14
Perceived Number of Contacts
Tijdsspanne: From day 0 to day 14

Among participants who completed the questionnaire: perceived number of contacts (too many, appropriate, too few).

From day 0 to day 14
Preferred Contact Mode for Future Follow-up
Tijdsspanne: From day 0 to day 14

Among participants who completed the questionnaire: preferred mode of contact for future follow-up questionnaires (categorical).

From day 0 to day 14
Characteristics of Responders Versus Non-Responders
Tijdsspanne: Day 14

Age, center, type of arthroplasty and time since surgery (when available in the host cohort) compared between participants who completed the questionnaire and those who did not, by arm (standardized differences).

Day 14
Agreement Between Voice-Captured Answers and Human Review
Tijdsspanne: From day 0 to day 14

In a random sample of completed voice interviews, agreement between the answers extracted by the AI agent and those coded by a reviewer listening to the recording, blinded to the extracted values (percentage agreement, kappa or intraclass correlation coefficient).

From day 0 to day 14

Andere uitkomstmaten

Uitkomstmaat
Maatregel Beschrijving
Tijdsspanne
Delivery of Email and SMS Invitations
Tijdsspanne: From day 0 to day 7

Proportion of emails and SMS delivered, bounced or filtered by the operator, as reported by the sending providers.

From day 0 to day 7
Outbound Call Outcomes
Tijdsspanne: From day 3 to day 7

Distribution of outbound AI call outcomes (answered, voicemail, no answer, filtered, failed) and call duration in the proactive AI voice calls arm.

From day 3 to day 7

Medewerkers en onderzoekers

Hier vindt u mensen en organisaties die betrokken zijn bij dit onderzoek.

Onderzoekers

  • Hoofdonderzoeker: Jules Descamps, MD, MSc, Hopital Lariboisière
  • Studie stoel: Geoffroy Nourissat, MD, PhD, Clinique des Maussins-Nollet, Ramsay Santé

Studie record data

Deze datums volgen de voortgang van het onderzoeksdossier en de samenvatting van de ingediende resultaten bij ClinicalTrials.gov. Studieverslagen en gerapporteerde resultaten worden beoordeeld door de National Library of Medicine (NLM) om er zeker van te zijn dat ze voldoen aan specifieke kwaliteitscontrolenormen voordat ze op de openbare website worden geplaatst.

Bestudeer belangrijke data

Studie start (Werkelijk)

9 november 2026

Primaire voltooiing (Geschat)

24 november 2026

Studie voltooiing (Geschat)

24 november 2026

Studieregistratiedata

Eerst ingediend

10 november 2026

Eerst ingediend dat voldeed aan de QC-criteria

10 november 2026

Eerst geplaatst (Werkelijk)

17 november 2026

Updates van studierecords

Laatste update geplaatst (Werkelijk)

17 november 2026

Laatste update ingediend die voldeed aan QC-criteria

10 november 2026

Laatst geverifieerd

1 november 2026

Meer informatie

Termen gerelateerd aan deze studie

Plan Individuele Deelnemersgegevens (IPD)

Bent u van plan om gegevens van individuele deelnemers (IPD) te delen?

NEE

Beschrijving IPD-plan

Identifiable data, call recordings and transcripts will not be shared. Aggregated results will be published.

Informatie over medicijnen en apparaten, studiedocumenten

Bestudeert een door de Amerikaanse FDA gereguleerd geneesmiddel

Nee

Bestudeert een door de Amerikaanse FDA gereguleerd apparaatproduct

Nee

Deze informatie is zonder wijzigingen rechtstreeks van de website clinicaltrials.gov gehaald. Als u verzoeken heeft om uw onderzoeksgegevens te wijzigen, te verwijderen of bij te werken, neem dan contact op met register@clinicaltrials.gov. Zodra er een wijziging wordt doorgevoerd op clinicaltrials.gov, wordt deze ook automatisch bijgewerkt op onze website .

Abonneren