Following a jury trial, appellant Jimi Dalton was convicted of unlawful possession with intent to distribute phencyclidine (POP), unlawful possession with intent to distribute cocaine, and unlawful possession of marijuana.1 On appeal, appellant challenges: (1) the trial court’s denial of the motion to suppress the drug evidence; (2) the trial court’s refusal to release the jury despite two deadlock notes and a Gallagher anti-deadlock instruction; (3) the trial court’s imposition of an allegedly vindictive sentence, which punished appellant for exercising his right to trial; and (4) the trial court’s refusal to conduct a Jencks Act2 inquiry regarding statements made by the testifying police officers related to the use of force investigation. We affirm the trial court’s ruling with respect to the first three issues, but remand with respect to the final issue for the trial court to conduct the requisite Jencks Act inquiry regarding statements by the testifying police officers related to the use of force investigation.

I.

On the evening of August 3, 2010, appellant Jimi Dalton was bicycling, and several officers of the Metropolitan Police Department’s mountain bike tactical unit were patrolling, near the 800 block of K Street in Northeast Washington, D.C. Upon seeing the police officers on bicycles behind him, appellant accelerated and, shortly thereafter, abandoned his bicycle in a traffic lane and ran onto the sidewalk. Appellant testified that the officers caused him *1009to stumble off his bicycle onto the sidewalk, where several officers beat him and handcuffed him to a fence. The police officers testified that appellant had his hands in his waistband while running onto the sidewalk and failed to respond to several police orders to show his hands. The police officers also testified that they physically struggled with appellant to place him under arrest and that, during the struggle, a black plastic bag which was later found to contain PCP, cocaine, and marijuana fell from appellant’s person. Although the parties disputed what led appellant to fall off his bicycle and the circumstances of his altercation with the police officers, it was undisputed that appellant was in a physical altercation with the police that caused appellant to suffer facial contusions and necessitated appellant being taken to the hospital. Because one of appellant’s bones was broken, his case was referred to the Metropolitan Police Department’s Force Investigation Team.3

At the pre-trial suppression hearing on December 17, 2010, appellant’s counsel learned that a use of force investigation was pending against at least one of the police officers in the instant case and requested a continuance in order to develop a “full record” before proceeding with witness testimony in the suppression hearing. Appellant’s counsel argued that the pending investigation report might be material to the defense and might have a bearing on the court’s understanding of the sequence of events, as well as the officers’ bias and credibility.4 However, the government stated that it had provided all discoverable material to appellant’s counsel and that there was no use of force investigation report. The trial court denied appellant’s request for material related to the use of force investigation because appellant did not have the right to delay the proceedings until the investigation was completed. On at least two occasions, the date of the suppression hearing and the morning before the presentation of evidence at trial, the court asked whether the government had provided Jencks material to appellant and accepted the government’s affirmative reply without any further inquiry. Subsequent to the court’s inquiries, but prior to the presentation of evidence at trial, the government informed the court that all four officers involved in appellant’s arrest had been subjects of the use of force investigation and that the government declined to further investigate or prosecute the officers. Appellant’s trial counsel then renewed his request for material related to the use of force investigation, but the trial court denied this request.5

During the pre-trial suppression hearing, appellant challenged the admission of the drug evidence. Previously, at the initial scheduling hearing, the court and appellant’s trial counsel agreed that the mo*1010tion to suppress would be dispositive. That is, if the suppression motion was denied, appellant would plead guilty and reserve his right to appeal the motion decision, but there would be no trial. The trial court reiterated its understanding of the dispositive nature of the motion at the close of the pre-trial suppression hearing. Following the testimony of two government witnesses (Metropolitan Police Department officers Richard Mazloom and Mohamed Ibrahim) and five defense witnesses (Christopher Young, Wilbert Atkins, Willie O’Neal, Vincent Gomillion, and appellant himself),6 the trial court found the government witnesses credible based upon their demeanor, their lack of embellishment (i.e., they testified that there was nothing suspicious about appellant prior to his flight), and the fact that appellant’s injuries were consistent with the officers’ account of events. Additionally, the trial court found that the testimony of appellant’s friend, Christopher Young, and appellant himself confirmed some details of the officers’ testimony. The court found it implausible that the officers would assault appellant, whom they did not know, without reason or merely because he was cycling with Young, someone in whom they were interested. The court also found it implausible that the officers would have knocked appellant off his bicycle without first telling or asking him to stop (as they did with Young), and noted that the officers did not strike or physically abuse Young. Finally, the court did not think that the officers’ knowledge of a pending use of force investigation regarding their conduct with appellant would give them motive or bias in testifying against him.

The trial court found that the defense witnesses were not credible based on their demeanor and other factors. Specifically, the court noted that O’Neal and Gomillion “used hostility toward the police,” and that their testimony contained factual inaccuracies (e.g., the direction of the officers’ chase and the location of the altercation). The court doubted whether O’Neal and Gomillion were present at the scene. Also, because O’Neal and Atkins testified that they were together the evening of appellant’s arrest, the court’s doubts about O’Neal affected its assessment of Atkins’s credibility. The trial court discounted the testimony of Young and Atkins “because of their friendship” with appellant, which gave them motive to testify in support of appellant. Finally, the court found appellant not credible based on his incentive to have his case thrown out and his admittedly recent use of PCP.7 The court concluded that appellant’s PCP usage affected his ability to accurately perceive and recall events and that it bolstered the officers’ testimony that they smelled PCP when they approached him. At the conclusion of the suppression hearing, the court denied appellant’s motion. However, appellant did not plead guilty; instead, he decided to proceed to trial.8

*1011The presentation of evidence at trial took approximately one day; the jury then deliberated for a number of hours over the course of three days. During that time, the jury sent several notes to the court (two notes regarding the evidence presented and two notes indicating that they had not reached agreement).9 After deliberating for four-and-a-half hours, the jury sent a note stating: “[w]e, the jury, are not all in agreement.” Appellant’s counsel requested a mistrial based upon jury deadlock, but the government noted that the jury did not use the word “deadlock.” The trial court decided to provide an initial jury instruction encouraging the jurors to deliberate further.10 After two-and-a-half additional hours of deliberation, the jury sent another note, stating, “[w]e, the jury, are deadlocked and feel that further deliberation would not lead to an agreement.” Appellant again requested a mistrial, but the trial court instead found it “reasonable” to provide a Gallagher anti-deadlock instruction.11 The jury returned to deliberate for another hour before being dismissed for the evening with the trial judge’s direction that they return the next morning. The next day, the jury deliberated for another two hours before returning a guilty verdict on all counts.

The trial court sentenced appellant to concurrent one year suspended sentences on each of the two counts of possession with intent to distribute, followed by three years of supervised release, also suspended, and one-year supervised probation with ninety days in a halfway house. Additionally, the court imposed a fíne of $1,000 for each of these two counts. On the possession of marijuana count, the trial court *1012sentenced appellant to a ninety-day suspended sentence, one year supervised probation, and a fine of $500. The trial court required the minimum payment into the crime victims fund on each count, which totaled $250.

At sentencing, the trial court cited appellant’s decision to “put on a parade of witnesses [at the suppression hearing] who ... clearly perjured themselves” without “even calling] them at trial” as a factor in its decision-making. The court stated that it did not consider appellant’s decision to “renege” on his agreement that the motion to suppress would be dispositive (i.e., that appellant would plead guilty if the motion was denied, rather than proceed to trial), or the fact that appellant did not testify truthfully (noting that appellant may not have remembered events clearly because he was high on PCP). The court also considered that, since his arrest, appellant had tested negative for drugs, that the amount of drugs confiscated upon his arrest was relatively small, and that appellant had a job and was working to support his wife and child.

Several months after appellant’s trial and sentencing were completed, on January 13, 2012, the Metropolitan Police Department (MPD) responded to appellant’s Freedom of Information Act (FOIA) request for records related to appellant’s allegation of excessive force. Among the records provided by the MPD were the Use of Force Incident Reports regarding the conduct of each of the four police officers concerned; each report was dated March 9, 2011 — the day that the police officers’ testimony and the trial concluded.

II.

A.

Appellant first contends that the trial court erred in denying his motion to suppress the drug evidence because the police unlawfully seized him through a show of authority, by rapidly pursuing him on their bicycles. Further, appellant contends that the police had no grounds to justify an investigative stop and that the drug evidence should have been suppressed as the fruit of an unconstitutional stop and seizure. In reviewing the trial court’s denial of the motion to suppress, we review factual findings for clear error and legal conclusions de novo. Beaner v. United States, 845 A.2d 525, 535 (D.C. 2004). Furthermore, “[i]t is the role of the trial court to assess the credibility of witnesses, and this court will not reverse a credibility finding unless it is clearly erroneous or lacking evidentiary support.” Bolanos v. United States, 938 A.2d 672, 685 (D.C. 2007) (citing Hill v. United States, 664 A.2d 347, 351 (D.C. 1995)).

1.

We first consider whether the police unlawfully seized appellant, in violation of the Fourth Amendment, when they pursued him on their bicycles. “A police officer may make a seizure by show of authority and without the use of physical force, but there is no seizure without actual submission; otherwise there is at most an attempted seizure, so far as the Fourth Amendment is concerned.” Plummer v. United States, 983 A.2d 323, 331 (D.C. 2009) (citing California v. Hodari D., 499 U.S. 621, 626 n. 2, 111 S.Ct. 1547, 113 L.Ed.2d 690 (1991)) (other citations omitted).

Here, two police officers deemed credible by the trial court testified that appellant accelerated his bicycling after looking back and noticing that two officers were behind him on bicycles, and that appellant subsequently abandoned his bicycle in a traffic lane and fled on foot onto the nearby sidewalk. Thus, even assum*1013ing that the police officers were attempting to seize appellant by pursuing him on bicycles, the fact that appellant continued to flee from them demonstrates that he had not submitted to their show of authority and was therefore not seized within the meaning of the Fourth Amendment until the officers physically struggled with and handcuffed him. See, e.g., Plummer, supra, 983 A.2d at 325 (concluding that defendant was not seized within the meaning of the Fourth Amendment when police officers approached him with their guns drawn and ordered him to show his hands because he did not comply with their show of authority). “ ‘[Submission’ under Ho-dari D. requires, at minimum, that a suspect manifest compliance with police orders.” Plummer, supra, 983 A.2d at 331 (internal quotation marks and citations omitted).

2.

We next consider whether the police had reasonable, articulable suspicion to stop appellant and, ultimately, whether the motion to suppress the drug evidence was properly denied. We review the trial court’s factual findings for clear error and its legal conclusions de novo. See Beaner, supra, 845 A.2d at 535.

“To justify an investigative stop, the police must be able to point to specific and articulable facts which, taken together with rational inferences from those facts, reasonably warrant that intrusion.” Plummer, supra, 983 A.2d at 330 (citations and internal quotation marks omitted). Here, appellant’s unprovoked flight, coupled with abandonment of his bicycle in the street and running onto the sidewalk, raised reasonable, articulable suspicion that appellant was engaged in wrongdoing, which justified the police officers’ investigative stop of appellant. See, e.g. Coghill v. United States, 982 A.2d 802, 808 (D.C. 2009) (concluding that appellant’s headlong flight into the woods and refusal to show his hands when repeatedly ordered to do so gave the police officer reasonable, artic-ulable suspicion sufficient for an investigative stop); Wilson v. United States, 802 A.2d 367, 370 (D.C. 2002) (concluding that appellant’s increased pace into an apartment building after seeing the police, coupled with his frantic pounding on an apartment door, was “an unprovoked instance of evasive behavior sufficient” to give rise to reasonable, articulable suspicion).

Here, as in Wilson, the grounds establishing reasonable, articulable suspicion to justify an investigative stop increased to probable cause to arrest appellant and search him incident to the arrest. See Wilson, supra, 802 A.2d at 372. In Wilson, an experienced police officer’s grounds for an investigative stop increased to probable cause when he recognized the tin foil packaging of PCP being removed from the pocket of appellant’s companion, as well as the strong odor of PCP coming from appellant. Id. In the present case, trained and experienced officers recognized the strong chemical odor of PCP upon nearing appellant; furthermore, appellant had his hands in his waistband and failed to respond to several police orders to show his hands. See Beaner, supra, 845 A.2d at 535 (“[T]he test for judging the existence of probable cause is whether a reasonably prudent police officer, considering the total circumstances confronting him and drawing from his experience, would be warranted in the belief that an offense has been or is being committed.”) (quotation marks and citation omitted). These circumstances, viewed together, warranted the officers’ belief that appellant was committing an offense and gave rise to probable cause for the officers to arrest and search appellant.

*1014“Once the police make a lawful arrest, they may search the arrestee’s person and the area within his immediate control, in order to prevent the arrestee from gaining possession of a weapon or destructible evidence.” Howard v. United States, 929 A.2d 839, 846 (D.C. 2007) (affirming the trial court’s denial of the suppression motion where police officers had reasonable, articulable suspicion to stop, and probable cause to arrest appellant, and found drugs on the sidewalk at appellant’s feet and on appellant’s person) (quotation marks and citation omitted). Here, the record reflects that the black plastic bag of drugs fell from appellant’s person as the police officers struggled to place appellant under arrest. We discern no error in the trial court’s conclusion that the police officers had reasonable, articulable suspicion to stop appellant and, subsequently, probable cause to arrest and search him. Because the black bag of drugs was recovered pursuant to appellant’s lawful arrest and search of the area within his immediate control, the trial court did not err in denying the motion to suppress.

B.

Appellant next contends that the trial court improperly coerced the jury into rendering a guilty verdict by giving the jury a Gallagher anti-deadlock instruction after the jury had already returned two deadlock notes to the court, and by ordering the jury to return for a third day of deliberations when the presentation of evidence had taken only one day. When the jury expresses that it is deadlocked, the trial court has discretion to determine which, if any, anti-deadlock instruction to give. Davis v. United States, 700 A.2d 229, 231 (D.C. 1997) (citing Epperson v. United States, 495 A.2d 1170, 1173 (D.C. 1985)). “It is ... an abuse of that discretion to give an anti-deadlock instruction under circumstances creating a substantial risk of juror coercion.” Hankins v. United States, 3 A.3d 356, 361 (D.C. 2010). We make two inquiries when reviewing allegedly coercive jury instructions. First, we consider “the inherent coercive potential of the situation before the court.” Ford v. United States, 759 A.2d 643, 647 (D.C. 2000) (quoting Harris v. United States, 622 A.2d 697, 701-02 (D.C. 1993)). Then we “examin[e] ... the actions of the trial judge in order to determine whether these actions exacerbated, alleviated, or were neutral with respect to coercive potential.” Id.

On this record, we cannot say that the trial court abused its discretion by first giving the standard jury instruction and then giving the Gallagher anti-deadlock instruction. See, e.g., Hankins, supra, 3 A.3d at 362 (stating that it is generally not coercive to give standard anti-deadlock instructions when the jury has deliberated for considerable time and declared itself unable to reach agreement); Davis, supra, 700 A.2d at 230-31 (concluding there was no abuse of discretion where the trial court gave the Winters instruction, which is more forceful than the Gallagher instruction, after the jury stated “that we are not going to reach a verdict no [matter] how long we sit here”). Furthermore, there is no indication in the record that the trial judge was singling out the minority of jurors responsible for the deadlock, or was even aware of which group was inclined to acquit or which group was inclined to find appellant guilty. See Ford, supra, 759 A.2d at 647-48 (suggesting that where a judge knows a juror is in the minority, giving an anti-deadlock instruction may single out the dissenting juror and coerce him/her into changing his/her mind). Therefore, we discern no abuse of discretion in the trial court’s decision to give the Gallagher instruction, and we cannot say *1015that the actions of the trial judge further exacerbated any inherent coercive potential.

C.

Appellant next alleges that the trial judge impermissibly increased his sentence beyond probation, to include ninety days in a halfway house and a fine of $2500, because appellant decided to exercise his right to a jury trial after he had indicated that the suppression hearing would be dispositive. We review “fundamental legal errors in the sentencing process ... de novo.” Thorne v. United States, 46 A.3d 1085, 1089 (D.C. 2012) (citing United States v. Rivera, 448 F.3d 82, 84 (1st Cir. 2006)). “In conducting that review, we must be satisfied that the defendant’s sentence reflects an individuated judgment as to the balance of deterrence and rehabilitation applicable in his case rather than a categorical approach of using a maximum or an increased sentence for a defendant who required the government to prove his guilt beyond a reasonable doubt.” Id. at 1089 (citation and internal quotation marks omitted). A trial judge may not punish a defendant for exercising his Sixth Amendment right to trial. Id. at 1090.

Here, during sentencing, the trial court commented that had appellant pled guilty prior to or following the hearing on the motion to suppress, “this would be an easy call for probation.” See, e.g., Thorne, supra, 46 A.3d at 1089-90 (noting that, while it is improper for a trial judge to increase a defendant’s sentence because he exercised his right to trial, it is permissible to withhold leniency that the judge might otherwise grant “in exchange for a defendant’s admission of wrongdoing”). However, the present case is distinguishable from Thome, where the trial judge repeatedly stated that appellant’s exercise of his right to cross-examine a government witness had “sentencing consequences,” and failed to mention any other factors deemed relevant by the government or the defense. See Thorne, supra, 46 A.3d at 1087-88. Additionally, in Thorne, the trial judge ignored both the government’s request for partial sentence suspension and the defense’s request for time served and probation, and sentenced defendant to the maximum sentence provided by law. Id. By contrast, here, the trial court indicated several factors upon which it based the increased sentence, including appellant’s decision to put on defense witnesses who, in the court’s view, testified falsely. The trial judge specifically noted that, “[w]hat trouble[d] [the trial court] the most, however, is that at the hearing on the motion to suppress [appellant] put on a .parade of witnesses who in [the court’s] view clearly perjured themselves, and [appellant] didn’t even call them at the trial.” Finally, unlike Thome, where the defendant faced one count of drug possession and received the maximum possible sentence, here, appellant faced not only one count of drug possession, but two additional counts of drug possession with intent to distribute and received a sentence considerably less than the maximum possible, which further tends to rebut the appearance of vindictiveness by the trial court in its sentencing of appellant. See Thorne, supra, 46 A.3d at 1086, 1088. Cf. German v. United States, 525 A.2d 596, 603 (D.C. 1987) (emphasizing that “the mere fact of a sentence increase does not show vindictiveness”).

Because the trial court conducted its sentencing based on relevant and individually tailored considerations, weighing appellant’s credibility and motivations during his testimony, as well as appellant’s lack of prior convictions and his family responsibilities, we conclude that the trial court was not vindictive in its sentencing and did *1016not punish appellant for exercising his Sixth Amendment right to a jury trial. Thus, we affirm the trial court’s sentence which included ninety days in a halfway house and a $2500 fíne.

D.

Finally, appellant contends that the trial court erred when it dismissed his requests for potential Jencks material — statements by the testifying police officers during the use of force investigation, which would have facilitated his cross-examination of the police officers’ credibility — rather than inquiring whether such statements existed and determining whether they were producible Jencks “statements.” Appellant contends that this error, coupled with the chilling impact of the trial court’s credibility findings regarding the defense witnesses at the pre-trial motion hearing, was not harmless.12

We afford “trial courts ‘considerable deference in ruling on Jencks Act issues,’ which we review for abuse of discretion.” Lazo v. United States, 54 A.3d 1221, 1231 (D.C. 2012) (quoting Johnson v. United States, 800 A.2d 696, 699 (D.C. 2002)). However, the “[trial] court must conduct a proper inquiry and make relevant findings” before this court “will defer to the trial court’s ultimate ruling on production.” Id. Any trial court error regarding application of the Jencks Act is then subject to a harmless error analysis. Id. at 1235 (citing Lyles v. United States, 879 A.2d 979, 982-83 (D.C. 2005)).

The Jencks Act is designed to facilitate the impeachment of testifying witnesses who have given “statements” to the government.13 See id. at 1231 (quoting Frye v. United States, 600 A.2d 808, 810 (D.C. 1991)). Under the Jencks Act, the trial court has an affirmative duty, upon motion of the defendant, to determine whether the requested material is in the government’s possession, whether the material is a “statement,” as defined by the Act, and whether the material relates to the subject matter of the witness’s testimony. See Lazo, supra, 54 A.3d at 1231-32 (quoting Bayer v. United States, 651 A.2d 308, 311 (D.C. 1994)) (citing Lyles v. United States, 879 A.2d 979, 983 n. 12 (D.C. 2005)).

In our recent decision in Lazo, we articulated the court’s duty to conduct an independent inquiry into potential Jencks material, including when that duty is triggered. See generally Lazo, supra, 54 A.3d 1221. Now, we clarify when a change in circumstances may require the trial court to conduct multiple Jencks inquiries regarding the same or related material. In Lazo, we held that to trigger the trial court’s duty, the moving defen*1017dant need not prove that the Jencks “statement” actually exists; rather, the defendant need only establish that there is reason to believe it exists. See id. at 1232 (citing Goldberg v. United States, 425 U.S. 94, 124, 96 S.Ct. 1338, 47 L.Ed.2d 603 (1976) (Powell, J., concurring)). The trial court must then determine, by interrogation or by in camera inspection, whether the requested material is in the government’s possession and whether it is a producible Jencks “statement.” See id. at 1232 (“It is only after the court has made the necessary factual determinations that it can decide whether disclosure is required.”). If the government produces the requested material, “the court must inspect [the material] in camera to determine whether [it] qualifies] as a Jencks ‘statement’ (assuming, of course, that the government does not concede the point).” Johnson v. United States, 800 A.2d 696, 701 (D.C. 2002). However, if the government is unable to produce the requested material, the trial court must conduct an evidentiary hearing to determine whether the material ever existed, and if applicable, the circumstances of the material’s loss or destruction. Id.

Here, appellant’s requests for potential Jencks statements related to the use of force investigation, at the suppression hearing and on the day before the presentation of evidence at trial, were sufficient to trigger the trial court’s duty to conduct Jencks inquiries. See Lazo, supra, 54 A.3d at 1232 (noting that “[t]he evidentiary proffer that triggers the court’s duty to inquire is not onerous”). At the pre-trial suppression hearing, when appellant first learned of the use of force investigation, appellant requested the investigation reports and related material,14 thus triggering the trial court’s duty to conduct a Jencks inquiry as to whether any potential Jencks “statements” related to the investigation existed. On that date, the government represented that “the police officer stated that there might be a use of force report pending, but ... there is no use of force report currently.” The trial court then asked whether there was Jencks material that the government had not turned over, to which the government replied in the negative. Directly following the trial court’s affirmation that appellant was “entitled to ... any statements that the officers made in connection with the report,” and appellant’s response that he had received no such material, the government represented that “[t]here’s nothing ... in relation to the incident, there’s nothing to get that already discovery hasn’t provided. There’s nothing in relation to the investigation.” Given that the use of force investigation itself was still pending, and in light of the court’s reminders regarding the government’s disclosure obligations and the government’s repeated assurances that no such undisclosed material existed, there was no reason to believe that such Jencks statements existed at that time.15 Cf. Lazo, supra, 54 A.3d at 1234 (concluding that where a witness had expressly testified that an interviewing police officer had taken notes on her statement, the trial court had erred in *1018“summarily accepting, without additional inquiry,” the government’s representation that no such notes existed). Thus, we cannot say that the trial court abused its discretion, at the time of the suppression hearing, in concluding that the government did not possess any undisclosed potential Jeneks statements related to the investigation.

However, on the day before the presentation of evidence at trial, when the government represented that the use of force investigation was complete,16 the trial court’s duty to conduct a Jeneks Act inquiry was triggered again by appellant’s renewed requests for any statements by the testifying police officers related to the use of force investigation.17 Given that the investigation was now complete, and given the government’s clarification that all four police officers had been subjects of the investigation, there was now reason to believe that potential Jeneks statements by the testifying police officers might exist and the trial court could not “summarily acceptQ without additional inquiry,” the government’s representation that no such undisclosed material existed.18 See Lazo, supra, 54 A.3d at 1234-35 (citations omitted) (affirming that the prosecutor’s lack of awareness of potential Jeneks statements, and the trial court’s reliance upon the prosecutor’s representation, did not supplant the court’s duty to “conduct such an inquiry as may be necessary”). Rather, upon appellant’s renewed requests, the trial court had the duty to independently determine whether potential Jeneks material — statements by the testifying police officers related to the use of force investigation — now existed, and whether such material was producible as a Jeneks “statement.” See, e.g., Johnson, supra, 800 A.2d at 699 (concluding that the trial court erred in accepting the prosecutor’s representation that there were no police notes of an interview with the complaining witness in the case, which there was reason to believe existed, and in not requiring the government to search for the notes and turn them over to defense counsel or for the court’s in camera inspection). Thus, on this latter date, the trial court erred when it declined to direct the government to search for, or otherwise inquire about, statements by the testifying police officers regarding the use of force investigation, the investigation reports themselves, or any other potential Jeneks material related to the investigation.

Had the trial court ordered the government to produce the potential Jeneks statements, and had the govern-*1019raent complied, the court could have then determined whether the material constituted Jencks “statements” within the meaning of the Act, and, if so, ordered disclosure. If the government had been unable to produce the potential Jencks statements, the trial court then would have had the duty to hold an evidentiary hearing to determine whether such statements existed, and if not, whether negligence or bad faith was involved in their unavailability.19 See Johnson, supra, 800 A.2d at 701. Additionally, if the potential Jencks statements were not made available, the trial court’s duty would have been to determine the extent to which appellant would be prejudiced by his inability to access the statements for cross-examination of the government witnesses. Id.

Finally, we determine whether the trial court’s error was harmless. See Lazo, supra, 54 A.3d at 1235. An error will be deemed harmless if we are assured “that the judgment was not substantially swayed by the error.” Id. (quoting Kotteakos v. United States, 328 U.S. 750, 765, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946)).

Here, as in Lazo, we “assume the worst” — that qualifying Jencks statements by the testifying police officers existed— and inquire whether we can nevertheless say that the error was harmless.20 Lazo, supra, 54 A.3d at 1237-38. As with the complaining witness in Lazo, here,, the police officers’ testimony was of primary importance to the government’s ease-in-chief because it established the factual chain of events that resulted in appellant’s arrest, the seizure of drug evidence, and appellant’s eventual conviction. Appellant presented a theory, bolstered by the testimony of defense witnesses, which challenged the government’s version of events. However, appellant had no Jencks material for impeachment purposes. The use of force investigation reports concerned the police officers who testified at the suppression hearing and at trial and the circumstances of appellant’s altercation with police and appellant’s arrest. Accordingly, the subject matter of these reports pertained to the police officers’ credibility, which was the linchpin of the government’s case against appellant. Consequently, statements by the testifying police officers related to the use of force investigation would likely have aided appellant to more effectively cross-examine the police officers at trial. See Lazo, supra, 54 A.3d at 1236-38 (remanding for an evidentiary inquiry into potential Jencks material even when defense counsel had other non-Jencks material for impeachment and there was substantial corroboration of the complaining witness’s testimony by five other witnesses); cf. Middleton v. United States, 401 A.2d 109, 123 (D.C. 1979) (concluding that error of not producing witness’s grand jury testimony was harmless in light of the “exceptionally strong case against appellant adduced from the testimony of the other witnesses”); Moore v. *1020United States, 657 A.2d 1148, 1152 (D.C. 1995) (concluding that there was “no reasonable likelihood that the police notes, whatever they may have contained, could have had a significant effect on the outcome of the trial” where defense counsel already had three Jencks statements for impeachment purposes).

Furthermore, the jury in the present case was deadlocked for a significant amount of time and the trial court recognized that the only issue for the jury’s deliberation was that of witness credibility.21 Hence, “assuming] the worst” — that there existed qualifying Jencks statements by the testifying police officers — and recognizing that any material that related to the officers’ credibility was critical to appellant’s defense, we cannot say “with fair assurance ... that the judgment was not substantially swayed by the error.” Lazo, supra, 54 A.3d at 1235, 1237 (quoting Kotteakos, supra, 328 U.S. at 765, 66 S.Ct. 1239). Thus, we conclude that remand is required.

Although we now know that material concerning the use of force investigation does exist, we do not know whether all potential Jencks statements were disclosed in response to appellant’s FOIA request. Consequently, we remand the case for the trial court to conduct the appropriate evidentiary inquiries, both to determine whether the disclosed use of force investigation reports contain qualifying Jencks “statements” that relate to the subject matter of the witnesses’ testimony, and whether other, perhaps undisclosed, qualifying Jencks statements exist. See Lazo, supra, 54 A.3d at 1231-32. If the trial court determines that qualifying Jencks statements exist, the trial court must then determine whether appellant was prejudiced at trial as a result of the nondisclosure. See Johnson, supra, 800 A.2d at 701. If the trial court determines on remand that appellant was prejudiced, appellant’s conviction must be vacated and appellant afforded a new trial.

Accordingly, we remand the ease to the trial court for this Jencks Act inquiry; in all other respects, we affirm the judgment of the trial court.

So ordered.