2024-07-22

086_ Development of a multivariable prediction model of recovery in schizophrenia

Research Question and Aims

The study builds on a previous umbrella review of putative predictors of outcomes across mental disorders, and aims to investigate whether, and to what extent, a literature-informed set of candidate predictors can build a valid prediction model of recovery in those with schizophrenia spectrum disorder. The PsyCourse dataset provides a unique opportunity to analyze and compare the effects of all these factors within a single model. A better understanding of the factors influencing the recovery and remission(s) in patients with schizophrenia could improve both treatment and diagnosis of psychosis.

Analytic Plan

Starting from a set of predictors (see below), a Bayesian adaptive sampling procedure will be selected to explore which specific set (including potential interactions) may produce the best fitting model to data. Once the model will be identified, a series of multiple linear regressions will be used. Moreover, considering the longitudinal nature of the dependent variable(s), multiple-waves and multigroup structural equation models will be used to assess if the influence of each predictor (and/or their interactions) may change over time. Finally, clustering methods (such as exploratory graph analysis), will be used to define potential profiles of the respondents. Such profiles will be used to assess potential differences in recovery. To enhance internal validity of the results, the cited models will be tested on a subsamples of data, to get a first set of estimates. Such estimates will be used to conduct power analysis studies, through simulations. Based on the results, the initial models will be tested on the remaining sample. Results will be posted on open repositories.
Dependent variable: Recovery of patients in the spectrum of schizophrenia is operationalized as global functioning (Yu et al., 2020), stability of the symptoms/minimum symptom severity plus improved social, educational or vocational attainment (Catalan et al., 2021). Remission of patients in the spectrum of schizophrenia is operationalized as stability of the symptoms and/or minimum symptom severity for at least 6 months according to the RSWG remission criteria (Andreasen et al., 2005; Catalan et al., 2021). Independent variables composing the model: Candidate predictors include: psychotic symptoms (both positive and negative) along with their global severity and frequency, duration of the untreated illness and medication adherence, presence of depressive symptoms, as well as quality of life and functioning of the patients. Additionally, the impact of drug, tobacco, and alcohol consumption, as well as the presence of comorbid conditions such as migraines, epilepsy, and diabetes, will be considered. Finally, the sex, age, education level and marital status of the subjects will be taken into account.

Resources needed

v1_id
v1_sex
v1_yob
v1_school
v1_prof_dgr
v1_ed_status
v1_age_1st_inpat_trm
v1_dur_illness
v1_1st_ep
v1_tms_daypat_outpat_trm
v1_cat_daypat_outpat_trm
v1_fam_hist
v1_diabetes
v1_epilepsy
v1_migraine
v1_ever_smkd
v1_age_smk
v1_alc_pst12_mths
v1_evr_ill_drg
v1_sti_cat_evr
v1_can_cat_evr
v1_opi_cat_evr
v1_kok_cat_evr
v1_hal_cat_evr
v1_inh_cat_evr
v1_tra_cat_evr
v1_var_cat_evr
v1_scid_dsm_dx
v1_scid_dsm_dx_cat
v1_scid_ever_halls
v1_scid_ever_delus
v1_scid_ever_psyc
v1_scid_age_fst_psyc
v1_scid_yr_fst_psyc
v1_nrpsy_mwtb
v1_cape_itm1A
v1_cape_itm1B
v1_cape_itm2A
v1_cape_itm2B
v1_cape_itm3A
v1_cape_itm3B
v1_cape_itm4A
v1_cape_itm4B
v1_cape_itm5A
v1_cape_itm5B
v1_cape_itm6A
v1_cape_itm6B
v1_cape_itm7A
v1_cape_itm7B
v1_cape_itm8A
v1_cape_itm8B
v1_cape_itm9A
v1_cape_itm9B
v1_cape_itm10A
v1_cape_itm10B
v1_cape_itm11A
v1_cape_itm11B
v1_cape_itm12A
v1_cape_itm12B
v1_cape_itm13A
v1_cape_itm13B
v1_cape_itm14A
v1_cape_itm14B
v1_cape_itm15A
v1_cape_itm15B
v1_cape_itm16A
v1_cape_itm16B
v1_cape_itm17A
v1_cape_itm17B
v1_cape_itm18A
v1_cape_itm18B
v1_cape_itm19A
v1_cape_itm19B
v1_cape_itm20A
v1_cape_itm20B
v1_cape_itm21A
v1_cape_itm21B
v1_cape_itm22A
v1_cape_itm22B
v1_cape_itm23A
v1_cape_itm23B
v1_cape_itm24A
v1_cape_itm24B
v1_cape_itm25A
v1_cape_itm25B
v1_cape_itm26A
v1_cape_itm26B
v1_cape_itm27A
v1_cape_itm27B
v1_cape_itm28A
v1_cape_itm28B
v1_cape_itm29A
v1_cape_itm29B
v1_cape_itm30A
v1_cape_itm30B
v1_cape_itm31A
v1_cape_itm31B
v1_cape_itm32A
v1_cape_itm32B
v1_cape_itm33A
v1_cape_itm33B
v1_cape_itm34A
v1_cape_itm34B
v1_cape_itm35A
v1_cape_itm35B
v1_cape_itm36A
v1_cape_itm36B
v1_cape_itm37A
v1_cape_itm37B
v1_cape_itm38A
v1_cape_itm38B
v1_cape_itm39A
v1_cape_itm39B
v1_cape_itm40A
v1_cape_itm40B
v1_cape_itm41A
v1_cape_itm41B
v1_cape_itm42A
v1_cape_itm42B
alc_5orm
Antidepressants
Antipsychotics
bdi2_sum
cgi_c
cgi_s
clin_fst_ill_ep_dur
clin_fst_ill_ep_hsp_dur
clin_fst_ill_ep_hsp
clin_fst_ill_ep_med_chg
clin_fst_ill_ep_oth_end
clin_fst_ill_ep_othr
clin_fst_ill_ep_psy
clin_fst_ill_ep_slf_end
clin_fst_ill_ep_suic
clin_fst_ill_ep_symp_wrs
clin_ill_ep_snc_lst
clin_sec_ill_ep_dur
clin_sec_ill_ep_hsp_dur
clin_sec_ill_ep_hsp
clin_sec_ill_ep_med_chg
clin_sec_ill_ep_oth_end
clin_sec_ill_ep_othr
clin_sec_ill_ep_psy
clin_sec_ill_ep_slf_end
clin_sec_ill_ep_suic
clin_sec_ill_ep_symp_wrs
gaf
idsc_sum
marital_stat
med_pst_sx_mths
med_pst_wk
Mood_stabilizers
no_cig
nrpsy_dg_sym
nrpsy_dgt_sp_bck
nrpsy_dgt_sp_frw
nrpsy_tmt_A_err
nrpsy_tmt_A_rt
nrpsy_tmt_B_err
nrpsy_tmt_B_rt
nrpsy_vlmt_check
nrpsy_vlmt_corr
nrpsy_vlmt_lss_d
nrpsy_vlmt_lss_t
nrpsy_vlmt_rec
Other_psychiatric
panss_sum_tot
sf12_itm0
sf12_itm1
sf12_itm10
sf12_itm11
sf12_itm12
sf12_itm2
sf12_itm3
sf12_itm4
sf12_itm5
sf12_itm6
sf12_itm7
sf12_itm8
sf12_itm9
Tranquilizers
whoqol_dom_env
whoqol_dom_glob
whoqol_dom_phys
whoqol_dom_psy
whoqol_dom_soc