Towards digitally-enhanced psychiatry

Digi­tal tech­no­lo­gy is alrea­dy trans­for­ming the world of heal­th­care, ope­ning up unpre­ce­den­ted pros­pects. From per­so­na­li­sing care to redu­cing hos­pi­tal over­crow­ding, the expec­ta­tions are asto­ni­shing. But for this revo­lu­tion to live up to its pro­mise, heal­th­care pro­fes­sio­nals need to both adopt these tools, while rethin­king how they are used. The i3-CRG labo­ra­to­ry, hea­ded by Étienne Min­vielle at École Poly­tech­nique (IP Paris), has laun­ched a series of semi­nars on the inte­gra­tion of digi­tal tech­no­lo­gy in heal­th­care. One of the most eager­ly awai­ted topics has been men­tal health, an area where tech­no­lo­gi­cal advances could real­ly change the game. Pro­fes­sors Pierre-Alexis Geof­froy and Jean-Bap­tiste Mas­son take a look back at the semi­nar ses­sion devo­ted to psy­chia­try, which was also atten­ded by Guillaume Couillard, Direc­tor Gene­ral of the Paris Psy­chia­try & Neu­ros­ciences GHU, and Raphaël Gaillard, Pro­fes­sor of Psy­chia­try at Uni­ver­si­té Paris-Cité.

“Given the eco­no­mic constraints on the sys­tem, it is impos­sible to ima­gine that eve­ryone will be going to hos­pi­tal in the future,” explains Pro­fes­sor Pierre-Alexis Geof­froy, a psy­chia­trist at GHU Paris. “One per­son told me that her son had deve­lo­ped sleep disor­ders quite ear­ly on, and that she had tried to have him moni­to­red by a child psy­chia­trist, to no avail. In the end, six years later, he was diag­no­sed with schi­zo­phre­nia. But child psy­chia­try will never be able to treat all chil­dren with sleep disor­ders or anxie­ty disor­ders. And so these digi­tal solu­tions, which are also less expen­sive, will be there to adapt to the level of inten­si­ty of care that needs to be put in place, but also to the level of inter­ven­tion that we can offer.”

“We talk about digi­tal heal­th­care as if it were science fic­tion,” admits Pierre-Alexis Geof­froy. “But the truth is that digi­tal tech­no­lo­gy is alrea­dy here, and we all use it on a dai­ly basis.” Today, 95% of prac­ti­tio­ners com­plete their patients’ files using digi­tal tools. “When we write pres­crip­tions, for example, AI can alrea­dy tell us about pos­sible drug inter­ac­tions,” adds the pro­fes­sor. “All of this is based on scien­ti­fic lite­ra­ture, which is upda­ted in real time.” The point, then, is not so much to take stock of the tools alrea­dy avai­lable and in use in the world of psy­chia­try. Rather, it’s about loo­king ahead to the pos­sible changes that this world will under­go as a result of the tech­no­lo­gi­cal advances that could be made.

“For a long time, we were trying to trans­pose various scales and assess­ments that we used in real life into the digi­tal world. But that didn’t work because adap­ting to new digi­tal tools requires us to rethink eve­ry­thing,” explains the psy­chia­trist. Exis­ting health appli­ca­tions are an example. Although fewer in num­ber than well­ness apps, only 15% of them fol­low a scien­ti­fic approach, mea­ning they are based on a stu­dy with proof of effec­ti­ve­ness. “When we look at using this type of solu­tion, we also rea­lise that there is a pro­blem with com­pliance,” he conti­nues. “Only 30% of people com­plete their pro­grammes. So, the ques­tion is : how do we deve­lop this type of solu­tion properly?”

“In men­tal health, we are for­tu­nate to have very robust models,” points out Pierre-Alexis Geof­froy. “To deve­lop this type of solu­tion in addic­to­lo­gy, for example, we need to adapt models with a known scien­ti­fic approach to digi­tal tech­no­lo­gy.” For example, an appli­ca­tion to encou­rage people to give up an addic­tion, such as smo­king, needs to be deve­lo­ped on the basis of an exis­ting model, such as Pro­chas­ka and Di Cle­men­te’s model of pre­pa­ra­tion for change. 

“If my solu­tion fol­lows the logic of ‘one size fits all’ we will esta­blish a balance of the pros and cons of what stop­ping smo­king will bring to the patient. Howe­ver, if the patient is alrea­dy in a relapse phase, this will not speak to him. They will need much more tan­gible pro­po­sals. Pro­chas­ka and Di Cle­men­te’s model is the­re­fore impor­tant for deter­mi­ning what phase the patient is in, and the­re­fore what type of fol­low-up they will need. If they are at the contem­pla­tion stage, they will need moti­va­tio­nal inter­views to try and cla­ri­fy with them what they want and what they are pre­pa­red to do. If they are alrea­dy in the action stage, we need to orga­nise with­dra­wal with them. And if they are at the relapse stage, we need to ask them about what they have alrea­dy done, to deter­mine what has wor­ked well and what has not. This is essen­tial to ensure that patients are com­mit­ted to the solu­tion and that they don’t stop eve­ry­thing after 5 minutes because the appli­ca­tion doesn’t meet their needs.”

Accor­ding to the pro­fes­sor, the pri­ma­ry inter­est of digi­tal tech­no­lo­gy lies in pro­vi­ding an addi­tio­nal ser­vice to the doc­tor’s prac­tice. Offe­ring patients more regu­lar moni­to­ring does not mean making them inde­pendent in their approach, or even asking them to be too invol­ved in this moni­to­ring. “Just under 10 years ago, the Monar­ca I¹ stu­dy had an inter­es­ting idea. Howe­ver, today, this solu­tion alrea­dy seems ‘has-been’,” he argues. “The idea was to self-moni­tor patients with bipo­lar disor­der, in order to pre­dict when they might relapse. To do this, the 61 patients were asked to record their symp­toms on depres­sion scales.” The authors of this stu­dy sho­wed that the more depres­sed patients were, the less they inter­ac­ted with the medi­cal team. Conver­se­ly, the more manic the patient, the grea­ter the num­ber and dura­tion of calls.

“The results were suf­fi­cient­ly clear for the patient’s condi­tion to be easi­ly clas­si­fied,” confirms Pierre-Alexis Geof­froy. “This led the authors to conclude that smart­phone appli­ca­tions were valid for real-time patient moni­to­ring.” Howe­ver, a second stu­dy fol­lo­wed this one, with far less favou­rable conclu­sions. “In this second stu­dy, the authors deci­ded to keep all the patients, even those who had stop­ped using the pro­po­sed solu­tion,” explains the pro­fes­sor. “The result was that there was no signi­fi­cant effect of self-moni­to­ring, and the authors even obser­ved that noting down depres­sive symp­toms eve­ry day wor­se­ned the patient’s men­tal state.” 

“This type of moni­to­ring is not inten­ded to replace the doc­tor, but rather to offer an addi­tio­nal ser­vice that he or she can pres­cribe.” So, it’s a new tool avai­lable to doc­tors to ensure that their treat­ment is effec­tive. Because, in addi­tion to the pos­si­bi­li­ty of remote patient moni­to­ring, digi­tal tech­no­lo­gy also offers the­ra­peu­tic solu­tions. “I often use the example of the­ra­py using aug­men­ted vir­tual rea­li­ty. I’m a psy­cho­the­ra­pist and I have a patient with a pho­bia of cockroaches. I can work with him on expo­sure to insects by pro­jec­ting insects around his hand using vir­tual reality.”

“Accep­tance of this type of tool is a major issue. When we ask stu­dents whe­ther they would pre­fer to be trea­ted using a digi­tal solu­tion or by a real per­son, only 9% of them choose the digi­tal option,” notes Pierre-Alexis Geof­froy. “We the­re­fore need to pro­vide evi­dence of the effec­ti­ve­ness of these methods to unlock these bar­riers. Digi­tal tech­no­lo­gy is not yet present in my prac­tice, because digi­tal solu­tions for pro­vi­ding real-time patient data are not yet avai­lable. I dream of one day, in my prac­tice as a psy­chia­trist, in addi­tion to my tra­di­tio­nal prac­tice, having digi­tal argu­ments to help me make deci­sions. So, I think that sup­port will be de fac­to grea­ter when such solu­tions, pro­ven to be effec­tive, are available.”

This type of moni­to­ring is not inten­ded to replace the doc­tor, but rather to offer an addi­tio­nal ser­vice that he or she can prescribe

Pro­fes­sor Jean-Bap­tiste Mas­son, a resear­cher at Ins­ti­tut Pas­teur, ques­tions the metho­do­lo­gies used to assess the effec­ti­ve­ness of digi­tal tech­no­lo­gy in medi­cal contexts. “At some point, if we want to prove that some­thing is effec­tive, we’ll have to car­ry out sta­tis­ti­cal tests,” he says. “We’ll have to com­pare one group with ano­ther, whe­reas the human mind is not easi­ly put into cate­go­ries. It is true that digi­tal tech­no­lo­gy pro­vides an enor­mous quan­ti­ty of data and enlarges the control group. In psy­chia­try, the lar­ger the group, the more hete­ro­ge­neous it will be. Sub-groups will emerge, and com­pa­ri­sons will become less spe­ci­fic. So, one dif­fi­cul­ty will be in trans­po­sing convin­cing results from a small sample of people to a lar­ger sample for which the results will be less reliable,” he adds. This limi­ta­tion is also reflec­ted in the num­ber of para­me­ters that can be mea­su­red digi­tal­ly : the more para­me­ters and data that are stu­died, the grea­ter the chance of cor­re­la­tions, due to ran­dom­ness, without them being signi­fi­cant.” These metho­do­lo­gi­cal limi­ta­tions are still hol­ding back the vali­da­tion and adop­tion of these digi­tal solu­tions by practitioners.

The mar­riage bet­ween digi­tal tech­no­lo­gy and men­tal health the­re­fore opens up fas­ci­na­ting pros­pects. But for these solu­tions to win the confi­dence of prac­ti­tio­ners and patients, they must be accom­pa­nied by solid proof of their effec­ti­ve­ness. And that’s not easy to do. As Pro­fes­sor Geof­froy sums up, “digi­tal tech­no­lo­gy is not inten­ded to replace the doc­tor, but it can become a valuable ally in our prac­tice.” With rigo­rous cli­ni­cal research and gra­dual adop­tion, digi­tal tech­no­lo­gy could well rede­fine the approach to psy­chia­tric care, making treat­ments more acces­sible, per­so­na­li­sed and effec­tive. Howe­ver, the final hurdles to the suc­cess­ful deve­lop­ment of these solu­tions will have to be overcome.

Pablo Andres