Epstein Data Wiki

Harry Fisch

Properties2
tagsperson, physician, urology, weill-cornell, mount-sinai, new-york, medical-network
aliasesHarry Fisch MD, Harry Fisch, M.D., Dr. Harry Fisch, Dr. Fisch, Harry Fish, Dr. Fish

59 min read

ProfessionPhysician. His own billing form is headed “HARRY FISCH, MD”1
Practice944 Park Avenue, New York, NY 10028. The address is printed on his own billing form and circulated by Epstein’s office in appointment notes12
FieldUrology and reproductive medicine. The titles he negotiated for in 2016 were “Professor of Urology” and “Director of Microsurgery (Urology)“3
E-mail[email protected], printed on the sheets and reachable as a search term4
Relationship to EpsteinTreating physician, correspondent and social guest. Epstein wrote to him on 9 May 2016: “harry i want to bank some sperm, where do I go=? prescription/ ?”5
Principal recordA floor of 1,112 documents reaching him under twenty-two search forms plus a hand-assigned residue, against 440 for the phrase “Harry Fisch”. He is the named sender in the first sender field of 86 of those 440, with a further 45 sheets declined because that field is covered or the header mapping is unreadable67
The document that defines itOn the night of 26 October 2016 Epstein drafted, in Fisch’s own first person, the counter-offer Fisch sent to Mount Sinai’s urology administration. Fisch replied: “I like your suggestions. Will send out now.”8

Harry Fisch is a New York urologist who treated Jeffrey Epstein, corresponded with him in both directions from at least December 2015 until 2019, and used him as an adviser and an introducer. The relationship is documented in the files released under the Epstein Files Transparency Act across three distinct registers at once, and the same message often carries two of them: clinical questions and answers, a running social life built around dinners with Woody Allen, and Epstein’s involvement in Fisch’s own career and business.

What the release documents most fully is the last of those three. Over nine months in 2016 and 2017 Epstein acted as Fisch’s negotiator against the Mount Sinai Health System, and the record of it is continuous. In September 2016 the hospital’s own concern about the “compliance risk” in hiring Fisch part-time was relayed to Epstein by Eva Andersson-Dubin, who told him “We should not get involved.”9 He got involved. In October he wrote out Fisch’s counter-offer for him, phrased throughout as if Fisch were speaking, down to “I this is satisfactory , please return a copy for signature”, and Fisch answered “I like your suggestions. Will send out now.”8 In November, when the hospital appeared to withdraw, Epstein drafted the protest letter and Fisch sent his own version of it to a named recipient at the hospital the same afternoon, copying Arthur Klein, the Mount Sinai figure Fisch’s own messages name as having agreed to cover his expenses.1011 In February 2017 Epstein read the contract itself and caught a substitution in it, and Fisch changed the document accordingly.12 In June 2017 Fisch reported the outcome to him: the expenses would not be paid, and “Tewari’s admin said we should consider no contract, b[u]t consider private practice at Sinai.”12

Nothing on this page is a finding of wrongdoing. Fisch did not respond to several requests for comment from the newspaper that first reported on these documents, and no statement by him has been located for this page.13

What the searches reach, and what they do not

The phrase "Harry Fisch" returns 440 documents in the full-text index, measured on 14 September 2026 as count(distinct efta_number) over pages_fts.text_content with no LIMIT. Our own published report Medical Professionals Investigation carried 588 for this subject; that figure does not reproduce, and the correction has been reported to the desk.14

440 is a floor, and a low one. Twenty-two search forms were run, each measured separately and unioned, and the union is 1,067. A further 45 documents were assigned to him by hand out of the residue of the bare token fisch. The floor stated on this page is therefore 1,112 documents, 2.53 times the phrase count.6

Four mechanisms put documents outside the phrase, and three of them are not reachable by any spelling generator.

The correspondents’ own shorthand. "dr fisch" returns 457 documents, 408 of them outside the base phrase, and it is the largest single arm on this subject. It adds 398 beyond the running union of the three arms run before it, which is the increment [^floor] records; the outside-base figure and the running-union increment are different quantities and neither stands for the other. A random sample of twenty-two drawn from the part lying outside every other arm was twenty-two of twenty-two this man: appointment notes, lab scripts, a cheque request, calendar reminders and dinner entries passing through Epstein’s office.15

A second spelling of the surname, used by the export and by people. The rendered chat transcripts of his correspondence with Epstein label him “Harry Fish”, which returns 154 documents, 152 of them outside the base phrase; "Dr Fish" returns a further 43 outside everything else, sampled fourteen of fourteen him.16 This is not only an export convention. The spelling was in doubt among Epstein’s own correspondents, and the release records them settling it. On 22 September 2016 at 20:01:50 UTC a participant whose sender label the producing party has covered wrote, in a thread whose other messages on the same sheet are labelled Woody Allen: “I am checking on Fish or Fisch… How do you spell his name?” At 00:36:26 UTC the next morning the sender label reads Self ( e:[email protected] ), which is Epstein’s own side of the export, and the message reads: “it is fisch. fyi”.1718

Scanner damage to the forename. "Harty Fisch" (12), "Hany Fisch" (11), "Hany Fish" (8), "Harry Fisc" (6), "arry Fisch" (5), "Hariy Fisch", "Harry Fiscn" and "Harry isch" (1 each) stand together on 30 documents outside the base phrase, and add 20 beyond the running union of the arms run before them. Those are two different quantities: 30 is the union of the eight arms less the base, and 20 is the sum of the eight increments [^floor] records. scripts/ocr_variants.py generated twenty-nine forms for this name and found six of these; the rest came from reading the residue.19

A misspelling by the writer of the document. "Harry Fischer" returns one document, a message asking that someone “text me Dr. Harry Fischer’s cell”; "Harry Fisck" returns one, a lunch entry in a day schedule. Neither is optical damage. Both were typed that way by a person, and no substitution table generates either.2021

The identifier arms. His e-mail handle harryfisch returns 28 documents and adds 3; his practice address "944 Park Ave" returns 48 and adds 4; the laboratory account code printed on the blood-work reports he ordered, C1720, returns 11 and adds 2; the phrase from his own covering message, "my sitcom treatment", returns 9 and adds 2. Two telephone arms were run and are not printed here as values: the office line that appears on his own billing form returns 10 documents and the mobile number his correspondents circulated returns 13, and neither adds a document the name arms had not already reached. The surname-first form "Fisch Harry" returns 15, which is how the clinical laboratory headed its reports, and adds 1.6

The arms that were run and rejected are as much a result as the arms that were kept, and they are named here so the next reader need not run them again. "Veru", the company he founded, returns 62 documents of which 45 lie outside the union; a sample of sixteen of those 45 was fifteen of sixteen a typist’s slip for the word “very” (“im very veru glad for you”, “works veru vent well”), and the arm was dropped. urologist returns 81 documents, and the outside-union remainder is dominated by a different urologist who appears in the same appointment book. madelon, the forename of the staffer in his office, returns 19, of which the outside-union share is mostly a Palm Beach property roll. clomid and clomiphene return 37 and 13, but they are a fact about Epstein’s medication rather than a key to this subject, and they are not counted toward the floor. HBRK reaches 35,112 documents and is Epstein-entity boilerplate, not an identifier for anybody.22

The enumeration, and its window. An fts5vocab prefix enumeration was built in the temp schema on a read-only connection, so nothing was written to the corpus. The f prefix range completed: 239,254 index terms, scanned exhaustively, filtered at Levenshtein distance 2 or less from fisch and fish over lengths 3 to 9. A whole-index run was attempted and did not complete: it was killed at ten minutes with no output, and no null is claimed from it. A separate probe generated all 105,338 strings within edit distance 2 of fisch and fish at lengths 3 to 8 and looked each up in the index; 5,439 of them exist as terms, and at that distance on a four- or five-letter word the candidate set is dominated by ordinary English (with, first, cash, wish), so it produced candidate strings and not findings. The forms that actually carry documents were found by reading, not by generating.23

What no method used here can reach. A person referred to only as “my urologist” or “the doctor”; an appointment booked in someone else’s name; a sheet that names only his office staff; a document naming only the laboratory or the radiology practice he sent work to; and any corruption of the first character of every one of his search forms at once. The number above is a floor with those four holes named in it, not a population.

Does he act, or is he only mentioned?

Every one of the 440 base documents was classified from its own first sheet, under a rule written down before it was run. Find the first sender label on that sheet, over the sixteen recorded label forms rather than the literal string From:; take that label’s value and no other; and where the value is covered, or the header mapping cannot be read, decline rather than guess. A sender header standing lower down the same sheet, inside a quoted ----Original Message-- block or a forwarded chain, is not a sender act, and neither is an alert line whose sender field is empty and whose name sits in the subject.

On that rule he is the named sender of 86 of the 440, 19.5 per cent. A further 140 documents address him in a To: or Cc: line without his being the sender. 169 name him in prose, in a day schedule or in a roster: iPhone and Apple Watch distribution lists, and his own billing forms read as a name block. On the remaining 45 the first sender field is covered or the header mapping is unreadable, and those are declined rather than assigned to either side. 86 + 140 + 169 + 45 = 440. 86 is therefore a floor and not a rate, because a covered sender field is neither an act nor a non-act. No automated mail was found in the sender bucket at all: the LinkedIn-style “on behalf of” pattern that has inflated other subjects’ sender counts on this wiki returns zero here.7

All 86 sender values were listed and read, and every one of them is a form of his name or his own e-mail address; nothing else stands in that field anywhere in the bucket. He writes to Epstein and Epstein writes to him, which is the anchor test passing in both directions.

The Mount Sinai term sheet

September 2016: the hospital’s concern reaches Epstein

On 27 September 2016 Eva Andersson-Dubin forwarded Epstein a chain that had started inside Mount Sinai. The innermost message is from Melissa C. Bellino, whose signature block on the sheet gives the title “Senior Director, Ambulatory Oncology & Th=rapeutic Infusion” and a mountsinai.org address, under the subject “REGARDING YOUR FRIENDS RECRUIT..”, and it relays an update from a third member of staff:

=r. Klein spoke to Dr. Tewari regarding the compliance risk in hiring Dr. F=sch part-time. They will set up a meeting with Fisch to discuss the approp=iate time and effort allocation in the coming week.

Epstein replied: “can you ask what =o they mean by compliance , his deal is microsurgery”. Andersson-Dubin’s own message sits at the top of the sheet:

cant have part time…that =s the problem Tewari will speak to him. We should not get involved.<=r>

The = runs are quoted-printable export damage on the face of the sheet and are reproduced as printed, not repaired. “Klein” is named in Fisch’s own analysis three weeks later as Arthur Klein. No sheet read for this page states the office either man held. A physician surnamed Tewari appears independently in Epstein’s own day schedules, where the entries book appointments with “Dr. Ash Tewari (Urologist)”; those entries record appointments and not attendances, and this page draws no identification between the two beyond the shared surname.924

October 2016: Epstein writes the counter-offer

The negotiation runs through a single thread, and the whole of it is Fisch consulting Epstein.

On 7 October 2016 at 10:08 am Fisch forwarded Epstein a message under the subject “Fwd: confidential—analysis you asked for.”3

On Tuesday 25 October Fisch sent Epstein his draft: “JE, below is what I would send to Urology administration.” It set out the offer as made, itemised, with his changes written in beside each line: base salary of $160,000 “Change to $250,000”; a clinical supplement of $400,000 with no change written against it; the administrative supplement, payable only on hitting a philanthropic target, “Change to 25% of all pilanthropy raised during the term of the contract”; total compensation of $585,000 “Change to $650,000 plus administrative supplement”. A second paragraph set out what he wanted on expenses:

According to Arthur Klein, Sinai would cover my office expenses as a stipend to me. I would personally pay all my expenses related to the office including staff. My annual expenses were $386k. I will need an additional medical assistant at $65k and an additional $50k for the ultrasonographer. Sinai would need to cover my office expenses at $500k.

The sheet prints the sentence beginning “Sinai would need to cover” against a grey highlight, as it prints each of the four “Change to” clauses. The misspelling “pilanthropy” is on the face of the sheet.3

Epstein answered at 7:04 am on 26 October: “Talk later needs edits”.3

Fisch revised and sent the result back the same day, under a produced header of 7:47:14 PM. The asks moved in both directions and the total rose. The base-salary ask fell, from $250,000 to $200,000. The clinical supplement, which the 25 October version had left standing at $400,000 with no change written against it, was now itself asked up, to $440,000, so that the combined base-and-clinical ask fell by $10,000, from $650,000 to $640,000. The administrative supplement, asked the day before as an open-ended 25 per cent of all philanthropy raised over the life of the contract, became a flat $40,000. And the stated total went from $650,000 plus an administrative supplement of no stated size to $680,000, a figure that now had the administrative supplement inside it.3

At 8:51 PM on 26 October, in a quoted header, Fisch wrote again:

Jeffrey, hope dinner with Woody went well. I just realized I can make Egg Creams with Lactaid milk. No one wants Woody to be gaseous.

Below is a new version without a hint of whining.

What do you think?

At 9:15 pm, in the same quoted clock, Epstein sent back the document Fisch would use. It is written in the first person singular, as Fisch:

I very much appreciate the offer , we are not very far apart , what i would suggest is the following.

and it ends:

I this is satisfactory , please return a copy for signature.

Between those two sentences it fixes the compensation at a base salary of “$200000”, a clinical supplement of $440,000 and an administrative supplement of $40,000 for a total of $680,000; converts the collections bonus to “50% of gross collections above $1,000,000”; and, against the same grey highlight the earlier sheet used, states “To be clear Mount Sinai has agreed to cover all office expenses.” and fixes the two titles.8

Fisch’s reply carries a produced header of 1:33:01 AM on 27 October 2016 and reads in full:

I like your suggestions. Will send out now. Thanks Harry

The produced Sent: fields on these sheets and the quoted headers inside them are on different clocks, so no interval across the two is stated here. The order is taken from the quoting structure, which preserves it, and it is not computed. That the produced field is UTC on this correspondent’s mail is measurable on his own thread: a message stamped Monday, January 4, 2016 6:59:23 PM in a produced header is rendered Jan 4, 2016 at 1:59 PM when another production quotes the same message, a difference of exactly the five hours that separated New York from UTC that week.25

November 2016: Epstein drafts the protest letter

Two weeks later the arrangement came apart, and the release holds both halves of the same afternoon.

On 9 November 2016 Epstein sent Fisch a message carrying an RFC date header of Wed, 09 Nov 2016 19:57:45 +0000. It opens with a pasted passage of newspaper commentary about the election held the day before, which has nothing to do with the subject, and then, below a gap, this:

dear. I would like to confirm our recent and frankly very perplexing conversation. I had agreed to your proposal and that of Arthur Klein, . salary , sharings , responsbilities. etc. Arthur and I were clear and had agreed that expenses would be in the form of a stipend. What I think I heard from you is that you no longer can stand by the proposed agreement. . If that is true, please confirm to me in writing. I am very surprised , and find it difficult to understand

The message has no subject line and the misspelling “responsbilities” is on the face of the sheet.10

At 3:48 PM the same day, in a quoted Outlook header, Fisch sent the following to “Parikh, Rishi”, copying “Klein, Arthur”, under the subject “Mount Sinai agreement”:

Dear Rishi

I would like to confirm my recent and frankly very perplexing conversation we had. I had agreed to your proposal and salary guarantee for 3 years. Arthur Klein was clear and we had agreed that all expenses would be covered in the form of a stipend. What I think I heard from you is that you no longer can stand by the proposed agreement. If that is true, please confirm to me in writing. I am very surprised, and find it difficult to understand.

Klein replied to the chain at November 9, 2016 at 5:25:35 PM EST with five words and a barred value: “Rishi, pls call me at”. Fisch forwarded the whole of it to Epstein under a produced header of 10:34 PM with a single word of his own, “FYI”.11

Read consistently, with the produced header as UTC and the two quoted headers as Eastern, the four messages fall inside one afternoon in that order. Each stamp is attributed above to the header it came from; no interval across them is asserted here.

February and June 2017: the contract, and the end of it

By 10 February 2017 there was a contract to mark up. At 12:43 PM Fisch sent it to Epstein from an unbarred address: “Take a look at the contract with my revisions. Let me know what y[o]u think before I send it back.” Four minutes later, on the same quoted clock, Epstein answered:

why salaey increase in place of expenses that w=s not the offer

At 2:18 pm Fisch replied:

You are correct. I changed the contract to inc=ude your suggestion (attached). Arthur Klein offered to pay the expen=es upfront so that I can pay them and avoid Sinai getting involved. H= recommended it and I like the idea.

Four and a half months later Epstein asked, at 7:46 AM on 26 June 2017, “yes, but what was the last communic=tion”. Fisch answered under a produced header of 12:25 PM:

Verbally I was told that Arthur Klein’= office will not handle this. Told to speak with Tewari directly. Tewari’s a=ministrator said that Klein’s office didn’t want to handle the expenses as t=ey previously agreed. Tewari’s admin said we should consider no contract, b=t consider private practice at Sinai.

The = runs are export damage on the face of the sheet and are reproduced as printed.12

What the release read for this page does not show is what happened after that. Five searches were written down before they were run and every returned document was opened: "Sinai term sheet" returns 3, all of them the October 2016 thread; "Director of Microsurgery" returns 8, all within the same negotiation family; "Professor of Urology" returns 50, of which the great majority are other people’s titles in unrelated material; "Arthur Klein" conjoined with Fisch returns 10, the latest of which is the June 2017 message above; and Tewari conjoined with Fisch returns 8, of which three are the September 2016 chain and its copies and the rest are appointment entries for a urologist of that surname. No countersigned instrument, no acceptance, no announcement and no later reference to either title held was returned by any of the five. That is a statement about those five searches and the 68 distinct documents they returned between them, and not a claim that the release holds no such record.26

Epstein as a patient

Fisch’s own billing form is in the release. It is headed with his name, his degree and the Park Avenue address, and the patient name written on it is Epstein’s. The service is recorded as an office visit, new patient, at the current procedural terminology code for a level II new-patient evaluation.1

The clinical correspondence runs alongside everything else and is often in the same message. On 31 December 2015 at 7:28 AM Epstein wrote: “is a headache a side effect of the clomophene? ,”. At 7:41 AM Fisch answered: “No. How long have you taken clomid?” At 8:07 AM Epstein replied: “4 days”. Four days after that, under a produced header of 6:59:23 PM on 4 January 2016, Fisch wrote the whole of the following message:

Any more headaches?

Harry

All four stamps in that chain are quoted headers except the last, which is the produced one, and the three quoted times stand in the order given.27

On 9 May 2016 at 2:36 PM Epstein wrote to him: “harry i want to bank some sperm, where do I go=? prescription/ ?”5

On 24 March 2016 Fisch closed a message about a television project with three clinical questions in one line: “Back pain gone? Schmeckel wor=ing? Did you gain more weight?”28

The administrative traffic around the care is heavier than the care. Epstein’s office arranged the appointments, chased the scripts, fetched the results and paid the bills: a message asking for a cheque to be written to Dr. Fisch for a day’s appointment, an instruction that “your lab script is in Dr. Fisch folder”, a note that another physician’s office was sending scan results across to him, a confirmation that the laboratory had his blood-work script, and a standing arrangement under which he and Andersson-Dubin were seen together at Epstein’s house rather than at either practice.2930

This page names no patient but Epstein, and prints no clinical particular for anyone else.

Referrals, in both directions

Epstein sent people to Fisch by name. On 10 December 2016 he wrote four words to a correspondent: “Go see dr Harry Fisch”. In another message he recommended him in terms: “I d like you to personally call dr harry fish. … he is the top guy in the us.”3132 A separate message circulating the practice’s details adds the address and the reason for the referral in one line.2

The traffic ran the other way too. Fisch forwarded Epstein curricula vitae under their own subject lines and asked, in February 2019, “Did you get the other names I sent you?”33 He forwarded him a newspaper obituary; he forwarded him a note on a state insurance law passed the week before; he forwarded him a medical-news item on a drug-labelling vote; and in February 2018 he wrote to Terje Rod-Larsen with Epstein copied.34

Woody Allen, and the dinners

Fisch appears repeatedly in Epstein’s dinner and screening arrangements, and the name that recurs beside his in them is Woody Allen’s.

The day schedules and calendar entries carry the pattern plainly. A dinner entry reads “7:00pm DINNER w/Woody, Soon Yi, Richard Axel and D[r] Fisch”; a recurring lunch entry puts him at the Second Avenue Deli; an internal note sets “dinner fisch, karp” inside a week’s planning list that names other guests in the same shorthand; and a line in Epstein’s own shorthand reads “would you like dinner with woody thurs? harry fish? i can also invite. ?”35

On 15 June 2017 Lesley Groff wrote to Woody Allen’s assistant about a screening the following day and gave the list of attendees: Jeffrey Epstein, Kathy Ruemmler, Harry Fisch. Allen’s assistant replied with the room’s limit and the entrance to use.36 In the chat transcripts Fisch thanks Epstein afterwards: “That was a great movie. I actually loved it. And it was in Woody’s screening room. Thanks again!!!”37

Epstein’s office also arranged his haircuts. One alert reads “Cancel Fisch cut at 12 tomorrow if I don’t hear from him”; a salon’s apology for an error at checkout is addressed to the office and concerns “H[arry] Fisch’s Haircut with Patrick today.”38

The stem-cell exchange, and Steven Victor

On 9 May 2017 a rendered iMessage transcript carries a short exchange in which a third physician’s name comes up. The sheet is one of the chat exports whose speaker attribution exists only on the image: a grey, left-aligned column under the printed label “Harry Fish” with a brown “HF” avatar, and a blue, right-aligned column with an orange avatar whose name label the producing party has covered. The sheet carries nine message bubbles, six grey and three blue, and all nine were checked against all four printed signals, bubble colour, alignment, printed label and avatar, before anything here was attributed. The four agree on each.

In the sheet’s own order and its own timestamps:

[label covered by the producer], 12:39:06 pm: Sultan said you are putting stem cells in Dick ??!!

Harry Fish, 12:41:07 pm: No. Not in Dick. He wants that but not me.

Harry Fish, 12:42:03 pm: That’s Steve victor

“That’s Steve victor” is Fisch’s message. Earlier versions of our own Medical Professionals Investigation gave it to Epstein; the correction was made in that report on 14 September 2026 and is confirmed here independently at 400 dpi. Two further bubbles stand between that message and the barred-label side’s answer at 12:42:23 pm, and they are not elided here: at 12:42:09 pm the barred-label side wrote “you will have fun with jonathan. a true character”, and the answer at 12:42:23 pm reads “VICTOR????! noooooo”.39

Steven Victor is a New York dermatologist and stem-cell entrepreneur with his own page here; that page cites this same sheet and reaches the same attribution. Three months earlier, on 8 February 2017, Epstein had forwarded Fisch a message of Victor’s with two words of his own, “can you believe?”40

The Nikolic introduction

On 1 November 2018 at 3:15 PM Epstein wrote a single line to two addresses. It has no subject, and its whole body is each man’s forename set against the other’s, twice, in the two orders. The second address was Boris Nikolic’s, the biotechnology investor who was named in Epstein’s will as a backup executor.

Nikolic replied at 4:13 pm, thanking Epstein, saying he had been glad to meet Fisch and asking Fisch to send him something about his company. Fisch answered at 5:34 PM with an investor deck attached. The message thanks Epstein for the introduction; offers the deck as the current one for his listed company and points at that company’s own site and at the site for its principal product; and then states the fact this page relies on, which is that his company had merged with the firm whose female condom it supplies to the public sector in Africa and Brazil through UNFPA and USAID, and which it had lately brought into the United States market largely through telemedicine. He signed it with his forename, and then with his name and degree.41

The record shows what came of it, in two stages, and the answer is no both times. On 3 November Nikolic read the deck and declined. He gave the reason as a restriction on his own fund, which he described as able to invest only in private companies; he said he expected to raise a fund late the following year that could hold listed companies as well; and he wrote that for the present he was unable to invest in a company such as Fisch’s.41

In the same message Nikolic pointed at a non-profit the Gates Foundation funds. Fisch took the opening and asked for a different thing: a list of ten countries and their unmet orders for the product, from Uganda to the Democratic Republic of the Congo, and then the ask itself, which is an introduction to the right person at the foundation, so that grants might be got to supply the device to every woman in Africa who asked for one.41

On 6 November 2018 Nikolic answered that too. He wrote that he was in Seattle that week and had met people from the foundation the evening before, and then:

I was told that BMGF will not fund (directly) female condom projects going forward. There were some projects several years back but these are now terminated. As I was suspecting, for the most part it is an issue of procurement so this is now in hands of partners such as PATH, PSI

Sorry that I can’t be of more help

All of it stands on one document, in the order given, and Epstein’s address is in the copy line of each of the three November messages that follow the introduction.41

The sitcom, and the stock tip

Fisch wrote a sitcom and used Epstein to move it. The covering message appears in nine documents: “Attached is my sitcom treatment written by me and 2 former Letterman w[r]iters. Let me know what you think.”42

On 24 March 2016 at 7:30 AM Epstein asked: “great= thanks, whre are you with the sitcom?” Fisch answered:

HBO. Terry Winter is interested and wa=ts to write the pilot. He’s looking for a writer. Terry created Boar=walk Empire, sopranos and Vinyl.

The = characters are on the face of the sheet.28

Four weeks later the answer changed, in a message that does two entirely different things in four lines. On 20 April 2016 at 8:20 AM Epstein wrote: “sitcom? co? whats new. = my tricygelerids went up to 750 so i stopped your miracl= cure”. Fisch replied at 1:11 PM, and the whole of his message reads:

Clomid is unlikely to cause a rise in t=iglycerides. But, good idea to stop. Let’s see what happens. Let’s hope you= back pain doesn’t come back.

Sitcom on hold. Busy on a road show re=arding my new company. Merger with FHCO scheduled for July.

Look at =HCO, price is low. Once we merge, will have a new name on Nasdaq.

Co=sider buying FHCO before we merge. That will help us. Proxy going out in a f=w weeks.

Harry

The ticker is FHCO, and two independent things establish it. The corpus text layer of the sheet quoted above renders the first occurrence as “Merger with RICO” and the third as “Co=sider buying NCO”; at 600 dpi that sheet prints FHCO in both places, and a second production of the same message, whose extraction is undamaged throughout, carries the three lines as “Merger with FHCO scheduled for July”, “Look at FHCO, price is low” and “Consider buying FHCO before we merge.” FHCO was the Nasdaq ticker of The Female Health Company, the firm with which Fisch’s own company merged, and which is the firm his message to Nikolic names as the one they had merged with, two and a half years later.4344

Below the quoted Epstein message on that sheet stands the confidentiality trailer Epstein’s own outgoing mail carried, which says of what it accompanies that it “may constitute inside information”. This page draws no conclusion from that juxtaposition, records no allegation, and states only what the sheet carries: a treating physician, in one message to his patient, answering a question about a medication and then telling him to buy a named listed security ahead of a merger, while on a road show for the company being merged, because “That will help us.”

What the record read for this page does not show is whether Epstein acted on it. The forms below were written down before they were run, and every returned document was opened. FHCO returns 4: three market-summary circulars of 2011 and 2013 that list the ticker inside an alphabetical table of others, and one of the two copies of Fisch’s message. The other copy is invisible to that query, because its extraction spells the ticker two different wrong ways, which is the same demonstration in miniature. "Female Health" returns 22, verupharma 11, "NASDAQ VERU" 16 and "Aspen Park" 1, and their members are the 2018 Nikolic thread and its copies, company material, and press. Nothing returned records a purchase, a holding, a transfer, an instruction to a broker, or any reference by Epstein to the ticker at all. That is a statement about those five searches and their returns, not about the release.45

Gifts

On 19 November 2018 an assistant to Epstein sent a message headed “GIFT from Jeffrey Epstein-Apple Watch Series 4” asking each recipient for a size, a case colour, a band and a shipping address. The distribution block on the first sheet names, among others, “Dr. Harry Fisch”, Martin Nowak, Rich Kahn, Terje Rod-Larsen, Dr. Bruce Moskowitz, Kathy Ruemmler, Eva Andersson, Cecile de Jongh and Darren Indyke. A separate roster in the same family is headed “2018 iPhone List” and carries his name in the same company.4647

What the record read for this page does not show

The documents do not establish that Fisch treated anyone Epstein is alleged to have trafficked, that he knew of the trafficking, or that any clinical decision of his was taken for a reason other than a clinical one. The correspondence read here concerns Epstein’s own care, Epstein’s referrals of named adults, and Fisch’s career and business. Where a message on these sheets concerns a third party’s health, this page neither names that person nor reproduces the particular.

On Mount Sinai, the release carries the negotiation from September 2016 to June 2017 and stops there. What it does not carry, on the five searches set out above, is any record of what was signed, of what Fisch did next, or of whether he took either title. On the sitcom, the record does carry the answer, four weeks after the HBO message: Fisch put it on hold for a merger road show. Each of those absences is scoped to named searches whose returns were opened, and each is stated as what was searched and what came back rather than as a claim about the release.

See also

Notes

Footnotes

  1. Billing form on the physician’s own printed stationery, headed with his name, degree and the Park Avenue address, with the patient name written in by hand and the service recorded as a new-patient office visit at the level II evaluation code. Read at the image at 400 dpi. https://epstein-data.com/EFTA00304570 p.1 A second copy of the same form, and a third sheet of the same family carrying a facsimile header, are at https://epstein-data.com/EFTA00304571 p.1 and https://epstein-data.com/EFTA00282980 p.1 ↩ ↩2 ↩3

  2. Referral note circulating the practice details in one line, including the street address and the reason for the referral; and a day-schedule appointment entry carrying the same address with a walking instruction. https://epstein-data.com/EFTA00587631 p.1 and https://epstein-data.com/HOUSE_OVERSIGHT_002254 p.1 Neither sheet was rendered for this page and the quotation of the address is taken from the text layer; the address is independently printed on the billing form at [^bill], which was read at the image. ↩ ↩2

  3. E-mail thread, Subject: Re: confidential--analysis you asked for. Produced header Sent: Wednesday, October 26, 2016 7:47:14 PM, To: [email protected], sender field barred, signed “Harry”. Quoted within it: From: jeffrey E. <[email protected]>, a To: line whose value is partly barred, Sent: Wed, Oct 26, 2016 7:04 am, body “Talk later needs edits”; and On Tuesday, 25 October 2016, followed by a barred sender, opening “JE, below is what I would send to Urology administration.” The innermost header carries From: Harry Fisch unbarred, with a barred value following it, To: Jeffrey Epstein <[email protected]>, Sent: Fri, Oct 7, 2016 10:08 am, Subject: Fwd: confidential--analysis you asked for. Both sheets read at the image at 400 dpi, 2026-09-14. The four “Change to” clauses and the sentence beginning “Sinai would need to cover” are set against a grey highlight on the sheet; “pilanthropy” and “$100,000/year” are as the image prints them, the second of which the text layer renders as S100.000/ ear. https://epstein-data.com/EFTA01738880 pp.1-2 ↩ ↩2 ↩3 ↩4 ↩5

  4. The address prints in the clear at the foot of the sheet at [^contract], in the line On Fri, Feb 10, 2017=at 12:43 PM, <[email protected]> wrote:, read at the image at 400 dpi on 2026-09-14. As a bare search token harryfisch returns 28 documents and as the tokenised address "harryfisch aol com" it returns 13. ↩

  5. E-mail, From: jeffrey E. <[email protected]>, Sent: Monday, May 9, 2016 2:36 PM, To: Harry Fisch, body “harry i want to bank some sperm, where do I go=? prescription/ ?” The = is export damage on the sheet as the corpus carries it; this sheet was not rendered and the quotation is from the text layer. https://epstein-data.com/EFTA02463679 p.1 ↩ ↩2

  6. All counts measured 2026-09-14 against full_text_corpus.db on a mode=ro connection, FTS5 only, one arm at a time, select rowid first and identifiers resolved in batches of 900, no LIMIT on any query, and every figure is count(distinct efta_number). The twenty-two arms and the documents each added beyond the running union, in the order run: "Harry Fisch" 440 (+440); "Harry Fish" 154 (+152); "Fisch Harry" 15 (+1); "dr fisch" 457 (+398); "Dr Fish" 62 (+43); "Hany Fisch" 11 (+6); "Hany Fish" 8 (+1); "Harty Fisch" 12 (+9); "Hariy Fisch" 1 (+1); "Harry Fisc" 6 (+2); "arry Fisch" 5 (+1); "Harry Fiscn" 1 (+0); "Harry isch" 1 (+0); "Harry Fisck" 1 (+1); "Harry Fischer" 1 (+1); harryfisch 28 (+3); the office-line arm 10 (+0); the cell arm 13 (+0); "944 Park Ave" 48 (+4); "944 Park Avenue" 13 (+0); C1720 11 (+2); "my sitcom treatment" 9 (+2). Union 1,067. The bare token fisch returns 974 documents, of which 100 lie outside that union; all 100 were read and classified individually, not sampled: 45 are this man, 53 are somebody or something else, and 2 could not be decided and are excluded. 1,067 + 45 = 1,112. The 53 rejected are a painter whose surname ends in l, a Los Angeles family in an intelligence report, a neuroimaging citation, a Klee artwork title recurring in an auction catalogue, four different men of the surname appearing in name lists, a swimwear company’s mailing list, two Florida appellate cases, a law firm, a bond-market columnist and an assortment of scanner noise. The two undecided are a bank statement line and a card-transaction line, each carrying the surname and nothing else. ↩ ↩2 ↩3

  7. Classification of all 440 base documents from each document’s own first sheet, re-run 2026-09-14 under a strict first-sender-field rule. The rule as it was applied, in order: take the lowest-numbered sheet of the document; scan its header region for sender labels anchored to line start over the sixteen recorded label forms; take the first such label, and its value only; decide the header layout from structure rather than from the presence of any inline value — a value on the label’s own line, a contiguous inline header block, a block of labels followed by a block of values mapped positionally, or an alternating label/value run — and return a refusal, not a value, where the first sender label’s value is covered, where the positional mapping fails a type check on the Sent/Date field, or where the line that would be taken as the value is body prose, a Bates stamp or a date. A header naming him anywhere else on the sheet, including inside a quoted ----Original Message-- or Begin forwarded message: block, is not a sender act. Result: sender field 86, addressed without being the sender 140, named in prose, day schedule or roster 169, declined 45. 86 + 140 + 169 + 45 = 440. The 86 is a floor: the 45 declined sheets are counted against neither side. Five of the 45 are calendar alert sheets of one family, on which the extraction carries no value against either the To: or the From: label and the only occurrence of the name is in the subject line or the alert text; one of the five was opened at the image and is described below, and the other four were read from the text layer alone, so no field state on them is asserted here. The desk’s own scripts/act_rate.py, run as act_rate.py '"Harry Fisch"' Fisch Fish with the keys as separate arguments — a pipe-joined key is refused by the tool and would score every class zero — returns SENDER-FIELD ACT 89 over the same 440 under a looser rule that accepts a Fisch header anywhere in the first sheet’s header region; its self-test passes. The two figures bracket the quantity and the stricter one is printed above. All 86 sender values were listed: every one is a form of his name or his own e-mail address. Three sheets were opened at the image at 400 dpi to check forms the extraction had damaged or mapped wrongly — https://epstein-data.com/EFTA01745449 p.1, whose sender field the extraction reads as Many Fisch and whose image prints Harry Fisch; https://epstein-data.com/EFTA02342405 p.1, whose first sender field is his own e-mail address rather than his name; and https://epstein-data.com/EFTA02209031 p.1, one of those five alert sheets, on which a single bar covers the To: and From: fields together and the name stands only in the subject line and the alert text. A scan of the sender bucket for the automated patterns that have inflated other subjects’ figures on this wiki, a no-reply or bounce address or an “on behalf of” construction, returned zero. ↩ ↩2

  8. E-mail, Subject: Re: Sinai term sheet. Produced header Sent: Thursday, October 27, 2016 1:33:01 AM, To: [email protected], sender field barred, body “I like your suggestions. Will send out now. / Thanks / Harry”. Quoted within it: From: jeffrey E. <[email protected]>, To: Harry Fisch followed by a bar, Sent: Wed, Oct 26, 2016 9:15 pm, and below that On Wed, Oct 26, 2016 at 8:51 PM, followed by a barred sender. The two sentences “To be clear Mount Sinai has agreed to cover all office expenses.” and the titles sentence are set against a grey highlight; the sheet prints “$200000” without a comma at the first bullet and “$440,000” and “$40,000” with them, and those are reproduced as printed. Read at the image at 400 dpi, 2026-09-14. https://epstein-data.com/EFTA01739370 pp.1-2 ↩ ↩2 ↩3

  9. E-mail, From: Eva Dubin with the address barred, Sent: Tuesday, September 27, 2016 2:01 PM, To: Jeffrey Epstein, Subject: RE: REGARDING YOUR FRIENDS RECRUIT.. Quoted within it: From: [email protected], Date: Tue,=27 Sep 2016 15:59:40 +0200, recipient barred; and an innermost message from Melissa C. Bellino, at a mountsinai.org address, under the subject REGARDING YOUR FRIENDS RECRUIT.. dated Tue, 2= Sep 2016 12:40:56 +0000, whose signature block gives the title “Senior Director, Ambulatory Oncology & Th=rapeutic Infusion”. The = runs are quoted-printable export damage on the face of the sheet and are not repaired. Read at the image at 400 dpi, 2026-09-14. A second copy of the same message is at https://epstein-data.com/EFTA02449379 and was not opened for this page. https://epstein-data.com/EFTA02348201 p.1 ↩ ↩2

  10. E-mail, From: "jeffrey E." <[email protected]>, To: Harry Fisch < followed by a bar >, an empty Subject: field, Date: Wed, 09 Nov 2016 19:57:45 +0000. The sheet carries a pasted passage of newspaper commentary about the previous day’s election, then a gap of several blank lines, then the draft quoted in the body. The sheet prints “responsbilities” and the spacing and stray full stops inside the draft are as printed. Read at the image at 400 dpi, 2026-09-14. https://epstein-data.com/EFTA01062841 p.1 ↩ ↩2

  11. E-mail, From: Harry Fisch < followed by a bar >, Sent: Wednesday, November 9, 2016 10:34 PM, To: Jeffrey Epstein, Subject: Fwd: Mount Sinai agreement, body “FYI / Sent from my iPhone / Begin forwarded message:”. Quoted below it: From: "Klein, Arthur" with the address barred, Date: November 9, 2016 at 5:25:35 PM EST, To: a barred value and "Parikh, Rishi" with the address barred, Subject: RE: Mount Sinai agreement, body “Rishi, pls call me at” followed by a barred value. No sheet read for this page states the office held by either named recipient. Below that: a sender field the producing party has covered, Sent: Wednesday, November 09, 2016 3:48 PM, To: Parikh, Rishi, Cc: a barred value and Klein, Arthur, Subject: Mount Sinai agreement, and the letter beginning “Dear Rishi”. The letter is signed “Sincerely,” and the name stands on the second sheet, which was rendered at the image in the gate pass on this page and carries his name and degree unbarred. Read at the image at 400 dpi, 2026-09-14. The two wordings separate cleanly in the index and were measured on 2026-09-14. Fisch’s sent letter, keyed on "salary guarantee for 3 years", returns 2 documents: this one and https://epstein-data.com/EFTA02670935 . Epstein’s draft, keyed on the word sharings, which appears in his version and not in Fisch’s, returns 3: https://epstein-data.com/EFTA01062841 at [^protest], https://epstein-data.com/EFTA02670937 and https://epstein-data.com/HOUSE_OVERSIGHT_031639 . The phrase common to both, "frankly very perplexing conversation", returns exactly those five. Only the two sheets cited at [^protest] and here were opened at the image. https://epstein-data.com/EFTA01783766 p.1 ↩ ↩2

  12. E-mail thread, Subject: Re: contract with revisions. Top-level: From: Harry Fisch < followed by a bar, Sent: Monday, June 26, 2017 12:25 PM, To: jeffrey E., and the message beginning “Verbally I was told”. Quoted below, in the sheet’s order: On Jun 26, 2017, at 7:46 AM, jeffrey E. <[email protected] and “yes, but what was the last communic=tion”; On=Mon, Jun 26, 2017 at 7:15 AM, followed by a barred sender, whose message is not on this sheet; From: harryfisch with the domain barred, To: jeevacation <[email protected]>, Sent: Fri, Feb 10, 2017 2:18 pm, and “You are correct.”; From: jeffrey E. <[email protected]>, To: Harry Fisch with a bar, Sent: Fri, Feb 10, 2017 12:47 pm, and “why salaey increase in place of expenses that w=s not the offer”; and On Fri, Feb 10, 2017=at 12:43 PM, <[email protected]> wrote:, whose message runs onto the second sheet and reads “Jeffrey / Take a look at the contract with my revisions. Let me know what y=u think before I send it back. / As always, / Thnaks / Harry / PS-look at how I spelled thanks. Am I dyslexic?” The = runs are quoted-printable export damage on the face of the sheet and are reproduced as printed. Sheet 1 read at the image at 400 dpi, 2026-09-14; sheet 2 read from the text layer only, and the quotation from it is marked as such here. Copies of the February 2017 exchange alone are at https://epstein-data.com/EFTA00692083 , https://epstein-data.com/EFTA01042740 and https://epstein-data.com/EFTA01042744 , none of which was opened. https://epstein-data.com/EFTA02385346 pp.1-2 ↩ ↩2 ↩3

  13. Gifts, Dinners, Deals: Justice Department Files Reveal Epstein’s Ties to Multiple Weill Cornell Doctors, Christina MacCorkle and Atticus Johnson, The Cornell Daily Sun, 19 February 2026, updated the same day. https://www.cornellsun.com/article/2026/02/gifts-dinners-deals-justice-department-files-reveal-epstein-s-ties-to-multiple-weill-cornell-doctors Fetched 2026-09-14, HTTP 200, 131,229 bytes. The article states, in its own words: “Fisch did not respond to several requests for comment.” It describes him as “a celebrity urologist who worked as a clinical professor of urology and reproductive medicine at Weill Cornell until 2021, according to his LinkedIn”, and states that “He also operated a private practice in Manhattan, where Epstein was a patient.” Four quotations were taken from this article for this page and each was verified character by character against the fetched text. ↩

  14. Medical Professionals Investigation, published on this project’s reports site, section 2.2. Its corpus figure of 588 for this subject does not reproduce: the phrase "Harry Fisch" returns 440 as count(distinct efta_number). The report has since been corrected to 440 and now also records the “Harry Fish” form. Three further corrections owed to that report were found in writing this page and have been reported to the desk. The report is a lead list and is not cited as a source for any claim on this page. ↩

  15. "dr fisch" returns 457 documents. 408 lie outside the base phrase; 398 is its increment over the running union at the point it was run; and 349 lie outside every other arm of the completed twenty. The sample below was drawn from the 390 that lay outside every arm run at that moment, which is a fourth quantity and is named here as such rather than printed as a total. Twenty-two were drawn at random from those 390 with a fixed seed and every one was read: all twenty-two concern this man. Three of the twenty-two carry the name only on their second sheet, which is why the sample was read across all sheets and not only the first. ↩

  16. "Dr Fish" returns 62 documents, 43 outside the union at the point it was run. Fourteen were drawn at random from the 43 and read: fourteen of fourteen this man, the returns being appointment requests made through Epstein’s office, a message reporting that “dr fish called me seeing him this am”, and instructions to send laboratory results directly. ↩

  17. Rendered chat transcript, printed p.2. The block iMessage / 2016-09-22 20:01:50 (UTC) [1] carries a Sender: line the producing party has covered with a black bar; the message beneath it reads “I am checking on Fish or Fisch… How do you spell his name?” No attribution of that message is made here. The five blocks immediately below it on the same sheet are labelled Sender: Woody Allen, unbarred, and they are stamped 2016-09-22 20:06:37 (UTC) [2], 20:06:37 (UTC) [1] carrying an image attachment, 20:06:37 (UTC) [1], 20:06:55 (UTC) [2] and 20:06:55 (UTC) [1]. Read at the image at 400 dpi, 2026-09-14, and the count re-read at the image on the same day in the fix pass, which corrected it from four. https://epstein-data.com/EFTA01212422 p.2 ↩

  18. Rendered chat transcript, printed p.4 of the same document; the sheet’s printed Bates is EFTA01212425. The block iMessage / [6] 2016-09-23 00:36:26 (UTC) prints Sender: Self ( e:[email protected] ) in the clear, and the message reads “it is fisch. fyi”. The corpus text layer gives the timestamp as 00:36:29; the image reads 00:36:26, and the image is followed here. Read at the image at 400 dpi, 2026-09-14. https://epstein-data.com/EFTA01212422 p.4 ↩

  19. python3 scripts/ocr_variants.py "Harry Fisch" --wide --count was run on 2026-09-14. It generated 29 forms, one from the desk’s measured glyph pairs and 28 generic single-edit forms, and returned non-zero page counts for arry Fisch, Harty Fisch, Harr Fisch, Harry isch, Harry Fish, Harry Fisc and Harry Fiscn. Its page figures are not document counts and are not reproduced here; each form was re-run as a document count before being unioned. Harr Fisch returns one document and is inside the union under another arm. The forms the generator did not reach, Hany Fisch, Hany Fish, Hariy Fisch, Harry Fisck and Harry Fischer, were found by reading the residue. ↩

  20. “Please ask ur team to text me Dr. Harry Fischer’s cell” Read from the text layer; not rendered. The sheet is a rendered chat transcript and no attribution is made from it here; it is cited only as the single return of that spelling. https://epstein-data.com/EFTA01616691 p.2 ↩

  21. Day-schedule lunch entry rendering the surname with a terminal k, at the Second Avenue Deli address, with a driving instruction. Read from the text layer; not rendered. https://epstein-data.com/HOUSE_OVERSIGHT_002684 p.1 ↩

  22. "Veru" 62 documents, 45 outside the union; sixteen of those 45 were drawn at random and read, and fifteen carry the string as a typist’s slip for “very”. urologist 81 documents, 63 outside the union; thirteen read, and the returns are dominated by appointment entries for a different urologist. madelon 19 documents, 5 outside the union; all five read, and four are a property roll. clomid 37 and clomiphene 13, whose outside-union share is Epstein’s correspondence with other physicians about the same medication and is not a key to this subject. HBRK 35,112. "Female Health Company" 19, 6 outside, not counted because the arm reaches the company and not the man. None of these six is included in the floor. ↩

  23. The fts5vocab table was created as temp.v on a mode=ro connection following CREATE VIRTUAL TABLE temp.v USING fts5vocab('main','pages_fts','row'), so no object was written to the corpus and there was nothing to drop. The f prefix range completed in 158 seconds over 239,254 index terms and was filtered at Levenshtein distance 2 or less from fisch or fish over lengths 3 to 9. A whole-index run across all thirty-six single-character prefixes was attempted and was killed at ten minutes having produced no output; no null is claimed from it, and the ranges it did not reach are unmeasured. Separately, all 105,338 strings within edit distance 2 of fisch or fish at lengths 3 to 8 were generated and looked up directly against the vocabulary; 5,439 exist as index terms, and the highest-frequency of them are ordinary English words, so the probe is recorded as a candidate generator whose output was not used. fts5vocab’s doc column counts rows, that is pages, and not documents; no figure on this page is taken from it. ↩

  24. Day-schedule entries in Epstein’s own appointment record carrying an appointment with a urologist of that surname at a Madison Avenue address, on sheets that carry no reference to Fisch. https://epstein-data.com/EFTA01785316 p.1 and https://epstein-data.com/HOUSE_OVERSIGHT_002251 p.1 Neither was rendered; the reading is from the text layer and no attribution of the September 2016 message to this person is made on that basis. The identification of “Tewari” in the Mount Sinai chain rests on the chain itself, which states only a surname. ↩

  25. The produced header at [^clomid] stamps the 4 January 2016 message Monday, January 4, 2016 6:59:23 PM. Three other productions of the same message quote it as On Mon, Jan 4, 2016 at 1:59 PM: https://epstein-data.com/EFTA00702875 , https://epstein-data.com/EFTA00837201 and https://epstein-data.com/EFTA01745449 . The five-hour difference is the offset between New York and UTC in the first week of January, which fixes the produced field on this correspondent’s mail as UTC and the quoted headers as local. None of the three quoting copies was rendered; the figure is from the text layer and is offered as corroboration of the zone, not as a quotation. ↩

  26. Five searches against the outcome of the negotiation, written down before they were run, measured 2026-09-14 as count(distinct efta_number), and every returned document opened at the text layer. "Sinai term sheet" 3: EFTA01739367, EFTA01739370 and EFTA02446492, all the October 2016 thread. "Director of Microsurgery" 8: the two sheets cited at [^sinai1] and [^sinai2] and their copies, plus EFTA01066449 p.8 and EFTA01616146 p.2, which carry Fisch’s own line “Sinai and be the director of microsurgery. This is exactly what I do now at NY Hospital.” "Professor of Urology" 50, of which the great majority carry other physicians’ titles in unrelated material. "Arthur Klein" conjoined with Fisch 10, the latest being the June 2017 message at [^contract] and its copies. Tewari conjoined with Fisch 8: EFTA02348201 and EFTA02449379, the September 2016 chain; EFTA02385346, the June 2017 message; and five appointment and scheduling entries for a urologist of that surname. Distinct union of the five: 68 documents. None is a countersigned instrument, an acceptance, an announcement, or a later reference to either title held. None of the 68 was rendered and no field state on any of them is asserted here. ↩

  27. E-mail, From: Harry Fisch unbarred, with a barred value following it, Subject: Re:, Sent: Monday, January 4, 2016 6:59:23 PM, To: "jeffrey E." <[email protected]>; body “Any more headaches? / Harry / Sent from my iPhone”. Quoted below, in the sheet’s order: On Dec 31, 2015, at 8:07 AM, jeffrey E. <[email protected]> wrote: / 4 days; On Thu, Dec 31, 2015 at 7:41 AM, Harry Fisch < followed by a bar > wrote: / No. How long have you taken clomid?; On Dec 31, 2015, at 7:28 AM, jeffrey E. <[email protected]> wrote: / is a headache a side effect of the clomophene? ,. The trailing comma is on the sheet. The extraction of this sheet renders one of the two quoted Epstein addresses as [email protected] and the other as [email protected]; both print as [email protected] at the image. Read at the image at 400 dpi, 2026-09-14. https://epstein-data.com/EFTA01738959 p.1 ↩

  28. E-mail, From: Harry Fisch with the address barred, Sent: Thursday, March 24, 2016 11A0 AM as the extraction renders it, To: Jeffrey E., Subject: Re: My sitcom. Body as the corpus carries it: “HBO. Terry Winter is interested and wa=ts to write the pilot. He’s looking for a writer. / Terry created Boar=walk Empire, sopranos and Vinyl.” followed by “Back pain gone? Schmeckel wor=ing? Did you gain more weight?” Quoted below: On Mar 24, 2016, at 7:30 AM, jeffrey=E. <[email protected]> and “great= thanks, whre are you with the sitcom?” The = runs are quoted-printable export damage and are reproduced as the row carries them; this sheet was not rendered, and the corrupted send-time string is quoted rather than repaired or converted. https://epstein-data.com/EFTA02348269 p.1 ↩ ↩2

  29. A cheque request for a day’s appointment, https://epstein-data.com/EFTA02046664 p.1 ; “Tell them your lab script is in Dr. Fisch folder”, https://epstein-data.com/EFTA00330105 p.1 ; a subject line recording another physician’s office sending scan results across, https://epstein-data.com/EFTA00321505 p.1 ; a laboratory confirming it held the script, https://epstein-data.com/EFTA02048946 p.1 ; blood-work reports headed with his name reversed and his laboratory account code, https://epstein-data.com/EFTA00599643 p.1 and https://epstein-data.com/EFTA00594095 p.1 . None of these six was rendered for this page and each is read from the text layer; no field state on any of them is asserted here. ↩

  30. Scheduling messages confirming a joint appointment at Epstein’s home. “Allison, we are confirmed for 1:30 pm at Jeffrey’s home with Eva, Dr. Fisch and Jeffrey” appears at https://epstein-data.com/EFTA02196373 p.1 ; day-schedule entries in the same family read “1:30 Appt w/Eva and Dr. Fisch”, https://epstein-data.com/EFTA02198085 p.1 and https://epstein-data.com/EFTA01060194 p.1 . Read from the text layer; not rendered. ↩

  31. E-mail, From: jeffrey E. <[email protected]>, Sent: Saturday, December 10, 2016 6:19 PM, Subject: Re:, body “Go see dr Harry Fisch”. The recipient is named in the header. Read from the text layer; not rendered. https://epstein-data.com/EFTA02668295 p.1 ↩

  32. “I d like you to personally call dr harry fish. [telephone number, not reproduced by this page]. he is the top guy in the us. he” Read from the text layer; not rendered. The sentence continues past the point quoted. https://epstein-data.com/EFTA01019263 p.1 ↩

  33. Two forwarded curricula vitae under their own subject lines, https://epstein-data.com/EFTA01029456 p.1 (From: Harry Fisch, To: Jeffrey Epstein <[email protected]>, Date: Thu, 14 Feb 2019 21:15:24 +0000, with an attachment named in the header) and https://epstein-data.com/EFTA02635005 p.1 (Sent: Thursday, February 21, 2019 12:04 AM, To: [email protected], body “Did you get the other names I sent you?”). Read from the text layer; not rendered. ↩

  34. Obituary forward, https://epstein-data.com/EFTA02629415 p.1 ; a note on a state insurance law, https://epstein-data.com/HOUSE_OVERSIGHT_019218 p.1 ; a drug-labelling item from a medical trade publication, https://epstein-data.com/EFTA02460598 p.1 ; and a message of 7 February 2018 addressed to Terje Rod-Larsen with Epstein in the copy line, https://epstein-data.com/EFTA02529021 p.1 . All four read from the text layer; none rendered. ↩

  35. Day-schedule dinner entry naming the director, his wife, a named scientist and “D[r] Fisch”, https://epstein-data.com/EFTA00459778 p.1 , with copies at https://epstein-data.com/EFTA00643092 and https://epstein-data.com/EFTA00667480 ; the recurring Second Avenue Deli lunch entry, https://epstein-data.com/EFTA02228761 p.1 ; “dinner fisch, karp” in Epstein’s own planning shorthand, https://epstein-data.com/EFTA02044764 p.1 , under a subject line that asks the date at https://epstein-data.com/EFTA00319875 p.1 ; and “would you like dinner with woody thurs? harry fish? i can also invite. ?” Each of these is read from the text layer and none was rendered. The extraction of the first renders “Dr” as “DE” and “Woody” as “Wbody” on two of the three copies, which is why the bracket is used. ↩

  36. Two messages of 15 June 2017 between Epstein’s assistant and the director’s assistant. From: Lesley Groff, Subject: Re: Screening Tomorrow, Date: Thu, 15 Jun 2017 18:02:29 +0000, body “Lauren the list of attendees is: / Jeffrey Epstein / Kathy Ruemmler / Harry Fisch / OK on your e” at https://epstein-data.com/EFTA00454411 p.1 ; the reply at Date: Thu, 15 Jun 2017 18:34:55 +0000, “Just keep in mind we can only have 10 total people in the room if he has any last m”, at https://epstein-data.com/EFTA00454421 p.1 , whose signature block gives the sender’s role. Read from the text layer; not rendered. Both quotations stop where the corpus row’s line breaks; neither sentence is complete on the row and neither is presented as complete. ↩

  37. Rendered chat transcript, printed p.4, date divider “2017/06/17”, under the printed label “Harry Fish”: “That was a great movie. I actually loved it. And it was in Woody’s screening room. Thanks again!!!” Read from the text layer of the same document whose printed p.3 was rendered at [^chat1]; this sheet was not itself rendered, and no attribution on it is asserted from the image. https://epstein-data.com/EFTA01616196 p.4 ↩

  38. “Alert - Cancel Fisch cut at 12 tomorrow if I don’t hear from him”, https://epstein-data.com/EFTA02209745 p.1 ; and a salon’s message to the office, “For H[arry] Fisch’s Haircut with Patrick today. Again, our apologies for the error at checkout.” at https://epstein-data.com/EFTA00448701 p.1 , whose second copy at https://epstein-data.com/EFTA00448704 extracts the forename as “Hasty” and the word “error” as “ertot”. Read from the text layer; not rendered. ↩

  39. Rendered chat transcript, printed p.3, date divider “2017/05/09”; the sheet’s last printed Bates is EFTA01616198, which is a page stamp inside this document and not a document identifier, and no corpus row exists under it. The sheet carries nine message bubbles. Six stand in a grey, left-aligned column under the printed label “Harry Fish” with a brown “HF” avatar; three stand in a blue, right-aligned column with an orange avatar under a label the producing party has covered with a black bar. All nine were checked against all four printed signals, bubble colour, alignment, printed label and avatar, and the four agree on each. The extraction of this sheet moves the right-hand bubbles out of position, so the order given in the body is the image’s and not the text layer’s. Read at the image at 400 dpi, 2026-09-14. https://epstein-data.com/EFTA01616196 p.3 ↩

  40. E-mail, From: jeffrey E. <[email protected]>, Sent: Wednesday, February 8, 2017 3:11 PM, To: Harry Fisch, Subject: Fwd: Huge Favor, body “can you believe?” The forwarded message below it is headed with the dermatologist’s name. This footnote records the same document and the same reading that the live page Steven Victor cites and states as read at the image at 400 dpi; it was not re-rendered for this page, and the reading is attributed to that page rather than claimed here. https://epstein-data.com/EFTA02661094 p.1 ↩

  41. E-mail thread on one document, four printed sheets, read at the image at 400 dpi on 2026-09-14 (pp.1-3). The header fields of this thread are described here in words and are not transcribed. A transcribed header line is a pasteable string belonging to the envelope rather than to the finding, and the finding here is what the messages say; every stamp relied on below is also stated in the body above, in words. Printed p.3 carries three messages. The lowest is Epstein’s, addressed to two recipients of whom Nikolic is one, sent on the afternoon of 1 November 2018 with an empty subject field and the four-word body. Above it stands Nikolic’s reply of the same afternoon, whose sender field is his own and which is addressed to Epstein and to Fisch together. Above both stands Fisch’s message of that early evening, sent from his own address to Nikolic, carrying the deck and its signature block. Printed p.2 carries Nikolic’s message of the early afternoon of Saturday 3 November, sent to Fisch at his own handle with Epstein’s address in the copy line, and the decline; the fund’s web address is printed on the sheet, and the corpus extraction of it differs from the sheet by one trailing character, so the sheet governs. Printed pp.1 and 2 carry Fisch’s reply with the ten-country list and the request for an introduction at the foundation. Printed p.1 carries Nikolic’s final answer, whose RFC date header is timed at 16:19 UTC on 6 November 2018, again with Epstein’s address in the copy line. The 1 and 3 November messages are summarised in the body; the 6 November message is summarised in its first sentence and quoted thereafter. https://epstein-data.com/EFTA01017442 pp.1-3 ↩ ↩2 ↩3 ↩4

  42. The covering message “Attached is my sitcom treatment written by me and 2 former Letterman w[r]iters. Let me know what you think.” appears in nine documents. The nine were reached by the phrase "my sitcom treatment" and the count is count(distinct efta_number); the clean copy of the string, without export damage, is at https://epstein-data.com/EFTA02385048 p.1 . Read from the text layer; not rendered. ↩

  43. E-mail, From: Harry Fisch unbarred, with a barred value following it, Sent: Wednesday, April 20, 2016 1:11 PM, To: jeffrey E., Subject: Re:. Quoted below it: On Apr 20, 2=16, at 8:20 AM, jeffrey E. <[email protected] <mailto:[email protected]> > wrote: and the message beginning “sitcom?”. Read at the image at 600 dpi, 2026-09-14, because the finding is three characters. The sheet prints FHCO at “Merger with FHCO scheduled for July” and at “Co=sider buying FHCO before we merge”, and prints =HCO at “Look at =HCO, price is low”, the = there being export damage on the face of the sheet. The corpus text layer of this sheet reads RICO for the first and NCO for the third, and neither string is on the image. All other = runs quoted here are likewise on the face and are not repaired. https://epstein-data.com/EFTA01772687 p.1 ↩

  44. A second production of the same message, Date: Wed, 20 Apr 2016 13:11:14 +0000, whose extraction carries no export damage: “Clomid is unlikely to cause a rise in triglycerides. But, good idea to stop. Let’s see what happens. Let’s hope your back pain doesn’t come back. / Sitcom on hold. Busy on a road show regarding my new company. Merger with FHCO scheduled for July. / Look at FHCO, price is low. Once we merge, will have a new name on Nasdaq. / Consider buying FHCO before we merge. That will help us. Proxy going out in a few weeks.” This copy was not rendered; it is used here only to corroborate the reading already taken from the image at [^tip], and no field state on it is asserted. Its RFC date header of 13:11:14 +0000 also corroborates, on a second thread, that the produced Sent: field of 1:11 PM on the sheet at [^tip] is UTC. https://epstein-data.com/EFTA00682560 p.1 ↩

  45. Five searches, written down before they were run, measured 2026-09-14 as count(distinct efta_number), every returned document opened at the text layer. FHCO 4: https://epstein-data.com/EFTA00682560 , the clean copy of the April 2016 message; and https://epstein-data.com/EFTA00386460 p.4, https://epstein-data.com/EFTA00433419 p.2 and https://epstein-data.com/EFTA02149039 p.4, which are morning market-summary circulars of 1 August 2013, 24 May 2011 and 30 January 2013 carrying the ticker inside an alphabetical list of others. https://epstein-data.com/EFTA01772687 is NOT among the four, because its extraction spells the ticker RICO and NCO; the image is what puts it in this population, and no query on the ticker would have reached it. "Female Health" 22, verupharma 11, "NASDAQ VERU" 16, "Aspen Park" 1; their members are the November 2018 Nikolic thread and its copies, company and press material. None of the returns records a purchase, a holding, a transfer, an instruction to a broker, or any reference by Epstein to the ticker. None of the returns was rendered except the sheet at [^tip]. ↩

  46. E-mail, Subject: RE: [External] GIFT from Jeffrey Epstein-Apple Watch Series 4, Date: Mon, 19 Nov 2018 15:33:02 +0000, with the originating message of Monday, November 19, 2018 6:50 AM quoted below it and its sender’s name barred, signed “Assistant to Jeffrey Epstein”. The distribution block spans the foot of the first sheet; the second sheet carries the questionnaire. Printed p.1 was read at the image at 400 dpi on 2026-09-14. The block prints his name in full, Dr. Harry Fisch; the corpus text layer breaks that name across a line and loses four characters, rendering it Dr. H. and Fisch on two lines, and the form the extraction produces is not what the sheet carries. On the same sheet the sender label of the 6:50 AM message stands above a bar where the name would be and a second bar over the address, which is why the sheet was opened: the extraction prints that label with nothing after it, and an absent value and a covered one are the same shape in the text layer. Printed p.2, which carries the questionnaire, was read from the text layer only, and no field state on it is asserted here. Two further copies in the same family are https://epstein-data.com/EFTA00485416 and https://epstein-data.com/EFTA00485440 . https://epstein-data.com/EFTA00485434 pp.1-2 ↩

  47. Roster headed “2018 iPhone List” carrying his name among ten others. Read from the text layer; not rendered. https://epstein-data.com/EFTA00285606 p.1 ↩

Backlinks