🏥 General Medicine — Research Topics

1. Pharmacist-Led Medication Reconciliation at Hospital Transitions of Care

• Why it's relevant: Polypharmacy errors at care transitions are a leading cause of preventable adverse drug events.
• Methodology: Retrospective cohort or prospective interventional study comparing pre/post pharmacist involvement.

2. Clinical Pharmacist Interventions in Antibiotic Stewardship Programs (ASP)

• Why it's relevant: ASP is a core PharmD competency and a high-priority public health issue.
• Methodology: Retrospective chart review comparing antibiotic use pre/post stewardship implementation.

3. Pharmacist Impact on Glycemic Control in Type 2 Diabetes (Ambulatory Care)

• Why it's relevant: Directly tied to evidence from CREDENCE and DAPA-CKD trials — high-impact, guideline-driven.
• Methodology: Randomized or quasi-experimental design in an outpatient clinic.

4. Deprescribing Potentially Inappropriate Medications (PIMs) in Elderly Patients

• Why it's relevant: Polypharmacy in elderly patients is a growing patient safety concern.
• Methodology: Prospective interventional study with pre/post medication reviews.

🧠 Psychiatry — Research Topics

1. Therapeutic Drug Monitoring (TDM) of Antipsychotics in First-Episode Psychosis

• Why it's relevant: Supported by the joint consensus of ASCP and AGNP on blood-level optimization (Management of First-Episode Psychosis, p.150).
• Methodology: Prospective observational study with TDM data and PANSS scoring.

2. Pharmacist Role in Monitoring and Managing Metabolic Side Effects of Antipsychotics

• Why it's relevant: Metabolic adverse effects are the leading cause of antipsychotic non-adherence.
• Methodology: Chart review + patient survey; pre/post intervention design.

3. Medication Adherence in Bipolar Disorder: Impact of Pharmacist Counseling

• Methodology: Prospective interventional study using Morisky Medication Adherence Scale (MMAS-8).

4. Polypharmacy and Drug-Drug Interactions in Patients on Long-Acting Injectable (LAI) Antipsychotics

• Methodology: Retrospective database/chart review using an interaction checker tool (Lexicomp, Micromedex).

🤖 Novel AI-Based Research Topics (PharmD Focus)

1. AI-Powered Clinical Decision Support for Pharmacist-Detected Drug Interactions

• Why it's novel: Goes beyond rule-based alerts; uses NLP and ML to prioritize alerts and reduce alert fatigue.
• Tools: Python (scikit-learn/XGBoost), hospital EHR data, FAERS database.

2. Natural Language Processing (NLP) for Adverse Drug Event Detection from Clinical Notes

• Why it's novel: Most ADEs go unreported through traditional systems; NLP enables passive surveillance.
• Tools: i2b2 NLP datasets, Python, HuggingFace Transformers.

3. AI-Based Prediction of Antipsychotic Treatment Response in Schizophrenia

• Why it's novel: Personalized antipsychotic selection remains largely trial-and-error; AI can enable precision psychiatry.
• Data sources: CATIE trial datasets, NIMH data repositories.

4. Large Language Model (LLM) Chatbot for Patient Medication Counseling: Accuracy and Safety Evaluation

• Why it's novel: Highly topical, publishable, and directly relevant to the future PharmD practice landscape.
• Methodology: Structured questionnaire design + expert panel review; no complex coding required — accessible for a 4th-year student.

5. Machine Learning to Predict 30-Day Hospital Readmissions in Polypharmacy Patients

• Why it's novel: Positions the pharmacist as a data-informed, outcomes-driven clinician.
• Data sources: MIMIC-IV (publicly available critical care database).

Summary Table

• General Medicine: Pharmacist-led deprescribing using STOPP/START in elderly inpatients
• Psychiatry: TDM of antipsychotics in first-episode psychosis
• AI-Based: LLM chatbot accuracy evaluation for medication counseling — no advanced coding required, highly relevant, and publishable in pharmacy practice journals

Novel PharmD Research Topics for 4th-Year Projects

🏥 GENERAL MEDICINE — Novel Topics

1. Pharmacist-Led Precision Deprescribing Using Pharmacogenomics in Polypharmacy Patients

• Research Question: Does PGx-guided deprescribing by clinical pharmacists reduce adverse drug events (ADEs) and hospitalizations compared to standard Beers Criteria–based deprescribing in adults ≥65 years?
• Why it's novel: PGx-guided deprescribing is an emerging intersection of pharmacogenomics and geriatric pharmacy — very few published studies exist.
• Methodology: Prospective quasi-experimental study; use CPIC guidelines + PGx panel results to identify high-risk drug-gene pairs before and after pharmacist intervention.
• Outcome measures: ADE rate, 30-day readmission, number of PIMs eliminated, patient quality of life (EQ-5D).

2. Gut Microbiome Modulation as an Adjunct to Pharmacotherapy in Metabolic Syndrome

• Research Question: Does adjunct probiotic/prebiotic therapy improve glycemic control (HbA1c), lipid parameters, and inflammatory markers (CRP, IL-6) in patients on standard pharmacotherapy for metabolic syndrome?
• Why it's novel: Backed by Kaplan & Sadock's data showing the "leaky gut hypothesis" in chronic inflammation — this bridges gastroenterology and metabolic medicine from a pharmacist's perspective.
• Methodology: Double-blind RCT; probiotic supplement (Lactobacillus + Bifidobacterium) vs. placebo added to standard metformin + statin therapy; 12-week follow-up.
• Outcome measures: HbA1c, fasting glucose, LDL, CRP, microbiome composition via 16S rRNA stool analysis.

3. Anti-Inflammatory Pharmacotherapy as an Adjunct in Treatment-Resistant Major Depressive Disorder (MDD) with Elevated CRP

• Research Question: In MDD patients with CRP > 3 mg/L who have failed ≥2 antidepressant trials, does adjunctive celecoxib vs. placebo (added to standard SSRI/SNRI) improve remission rates at 8 weeks?
• Why it's novel: Kaplan & Sadock (p. 5156) confirms: "Evidence that proinflammatory cytokines such as IL-1 and IL-6 are frequently elevated in depression has stimulated interest in the potential role of anti-inflammatory agents as antidepressants." No consensus has emerged yet — a gap ideal for PharmD research.
• Methodology: Prospective, double-blind, placebo-controlled RCT; stratify by baseline CRP; PHQ-9/MADRS scoring.
• Outcome measures: MADRS remission rates, CRP levels, side effect profile (GI, renal), quality of life.

4. Pharmacist Intervention on SGLT2 Inhibitor Utilization Gaps in Heart Failure with Reduced Ejection Fraction (HFrEF)

• Research Question: Does a structured pharmacist-led educational intervention targeting both prescribers and patients improve SGLT2 inhibitor initiation rates, adherence, and clinical outcomes (NYHA class, NT-proBNP) in HFrEF?
• Why it's novel: Focuses on the pharmacist as an implementation science driver — bridging evidence-to-practice gaps in cardiovascular pharmacotherapy.
• Methodology: Pre-post interventional study in a cardiology outpatient clinic; pharmacist delivers individualized education, monitors for UTI/DKA, counsels on sick-day rules.
• Outcome measures: SGLT2i initiation rate, adherence (PDC ≥80%), 6-month HF hospitalization rate, patient-reported outcomes.

🧠 PSYCHIATRY — Novel Topics

5. Gut Microbiome–Psychiatry Axis: Probiotic Supplementation as Adjunct Therapy in First-Episode Schizophrenia

• Research Question: Does probiotic supplementation (Lactobacillus rhamnosus + Bifidobacterium longum) added to antipsychotic therapy in FEP patients reduce negative symptom severity, antipsychotic-induced weight gain, and bowel dysfunction over 16 weeks?
• Why it's novel: Kaplan & Sadock (p. 695–696) states: "Reduced prevalence of Lactobacillus and Bifidobacterium species were noted among subjects with first-episode psychosis… differences in microbiota predicted severity of negative symptoms." Direct pharmacist-relevant intervention with no established standard of care.
• Methodology: Double-blind RCT; PANSS scoring for symptom severity; metabolic monitoring (weight, BMI, glucose, lipids).
• Outcome measures: PANSS negative subscale score, BMI change, bowel symptom score, IL-6/CRP inflammatory markers.

6. Therapeutic Drug Monitoring (TDM) + Pharmacogenomics in Optimizing Antipsychotic Dosing in Treatment-Resistant Schizophrenia

• Research Question: Does a pharmacist-driven protocol combining routine TDM with CYP2D6/CYP1A2 PGx testing improve clinical response (PANSS total score), reduce side effects (EPS, metabolic syndrome), and lower dose adjustments compared to standard dosing in treatment-resistant schizophrenia?
• Why it's novel: TDM alone is underutilized; adding PGx represents the next frontier in antipsychotic personalization — supported by the Maudsley Prescribing Guidelines (15th ed.) and the ASCP/AGNP joint consensus.
• Methodology: Prospective observational or quasi-experimental study in an inpatient/outpatient psychiatric unit; pharmacist performs TDM + PGx-guided dose optimization.
• Outcome measures: PANSS score reduction, plasma drug levels (within vs. outside therapeutic window), metabolic parameter changes, re-hospitalization rate.

7. Psilocybin-Assisted Pharmacotherapy for Treatment-Resistant Depression (TRD): A Systematic Review and Meta-Analysis

• Research Question: What is the comparative efficacy and safety of psilocybin vs. conventional antidepressants (SSRIs/SNRIs) in treatment-resistant MDD, based on current RCT evidence?
• Why it's novel: Kaplan & Sadock (p. 5156): "In a recent double-blind, randomized controlled trial psilocybin was comparable to escitalopram in reducing depression." Psilocybin is approaching FDA breakthrough therapy status — this is a high-impact, publishable topic for a PharmD.
• Methodology: Systematic review + meta-analysis (PRISMA guidelines); databases: PubMed, Cochrane, ClinicalTrials.gov; outcomes: MADRS/HAM-D remission rates, adverse events, relapse rates.
• No patient recruitment needed — ideal if clinical access is limited.

8. Digital Therapeutics (DTx) for Medication Adherence in Bipolar Disorder: A Pharmacist-Led Implementation Study

• Research Question: Does pharmacist-supervised use of a digital therapeutic app (symptom tracking + medication reminders + psychoeducation modules) improve adherence to mood stabilizers (lithium/valproate/lamotrigine) and reduce relapse rates in Bipolar I/II patients?
• Why it's novel: Bridges digital health, clinical pharmacy, and psychiatry. Kaplan & Sadock (p. 4218) highlights "text-to-quit programs, apps, and digital therapies" as emerging areas; this extends the concept to bipolar pharmacotherapy adherence.
• Methodology: Prospective RCT; pharmacist delivers app-based coaching; Morisky MMAS-8 for adherence; YMRS/MADRS for mood assessment.
• Outcome measures: Medication adherence rate, mood episode recurrence rate, psychiatric emergency visits, patient satisfaction (mHealth App Usability Questionnaire).

Summary Table

Top Picks for a PharmD Student

Sources: Kaplan & Sadock's Comprehensive Textbook of Psychiatry (p. 695–696, 5156); The Maudsley Prescribing Guidelines in Psychiatry, 15th ed.; Stahl's Essential Psychopharmacology; Management of First-Episode Psychosis and Schizophrenia (p. 150)

AI-Integrated Research Topics for PharmD 4th-Year Projects

🏥 GENERAL MEDICINE — AI-Based Topics

1. Using Machine Learning to Predict Adverse Drug Reactions (ADRs) from EHR Data

• Research Question: Can an ML model predict ADR occurrence in hospitalized patients with ≥5 medications with higher accuracy than standard clinical screening tools?
• Why it's novel: Harrison's (2025, p. Machine Learning chapter) confirms that "ML models perform on par with expert physicians on tasks ranging from image interpretation to predicting hospital readmission from EHR notes." Applying this to pharmacovigilance is a direct PharmD extension.
• Tools you'll use: Python (scikit-learn, XGBoost), hospital EHR data or publicly available MIMIC-IV database, FAERS (FDA Adverse Event Reporting System).
• Outcome: Sensitivity, specificity, AUC-ROC of the model vs. standard Beers Criteria screening.
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★

2. NLP-Based Automated Extraction of Drug-Drug Interaction Alerts from Physician Progress Notes

• Research Question: Does an NLP model detect clinically significant DDIs in unstructured EHR notes that are missed by rule-based pharmacy alert systems, and what is its precision/recall vs. pharmacist manual review?
• Why it's novel: Harrison's (2025) notes that "text-processing ML models have been successfully applied to analyze physician notes in the EHR and predict hospital readmission." Extending this to DDI detection is a critical pharmacy safety application.
• Tools you'll use: Python, HuggingFace Transformers (BioBERT/ClinicalBERT), i2b2 NLP clinical dataset or hospital EHR notes (de-identified).
• Outcome: Precision, recall, F1-score of NLP model vs. Lexicomp/Micromedex alerts.
• Difficulty: ⭐⭐⭐⭐ | Novelty: ★★★★★

3. Large Language Model (LLM) Evaluation for Medication Counseling Accuracy in Chronic Disease

• Research Question: How accurate and safe are LLM-generated medication counseling responses for top 20 chronic disease drugs compared to pharmacist-verified gold-standard answers? What categories of errors are most common?
• Why it's novel: Harrison's (2025) acknowledges that "LLMs have demonstrated the ability to excel at medical professional competency examinations and engage in diagnostic reasoning." However, accuracy in pharmacy-specific counseling has not been robustly evaluated.
• Tools you'll use: ChatGPT API / Gemini / Claude, standardized questionnaire design, expert pharmacist panel for blinded scoring (5-point Likert accuracy scale). No coding required — accessible for all PharmD students.
• Outcome: Accuracy score per drug class, error taxonomy (wrong dose, wrong frequency, missed contraindication, missed interaction).
• Difficulty: ⭐⭐ | Novelty: ★★★★★

4. AI-Powered Clinical Decision Support Tool for Antibiotic Selection in Community-Acquired Pneumonia (CAP)

• Research Question: Does an ML-based antibiotic CDSS improve guideline-concordant antibiotic prescribing rates, time-to-appropriate therapy, and clinical outcomes (length of stay, 30-day mortality) in CAP patients compared to standard physician prescribing?
• Why it's novel: Antibiotic stewardship + AI is a rapidly growing field with direct PharmD relevance; combines pharmacokinetics/pharmacodynamics with predictive modeling.
• Tools you'll use: Python (logistic regression, decision tree, Random Forest), local hospital antibiogram data, IDSA/ATS CAP guidelines as the benchmark.
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★

5. Predicting 30-Day Hospital Readmission in Polypharmacy Patients Using Machine Learning

• Research Question: Which ML model (logistic regression vs. random forest vs. gradient boosting) best predicts 30-day readmission in polypharmacy patients, and which drug-related problem features (ADR, non-adherence, under-dosing) contribute most?
• Tools you'll use: MIMIC-IV (free publicly available ICU database), Python/R, SHAP values for feature importance (explainability).
• Outcome: AUC-ROC, top predictive drug-related features, actionable pharmacist intervention targets.
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★

🧠 PSYCHIATRY — AI-Based Topics

6. AI-Based Prediction of Antipsychotic Treatment Response in First-Episode Schizophrenia

• Research Question: Can an ML model trained on clinical and biological baseline features predict antipsychotic response (defined as ≥30% PANSS reduction at 12 weeks) better than clinician judgment alone?
• Why it's novel: Treatment selection in schizophrenia remains trial-and-error. ASCP/AGNP consensus supports optimization — AI pushes this toward true precision psychiatry.
• Tools you'll use: Python, CATIE trial dataset (publicly available), PANSS scoring data, SHAP for explainability.
• Outcome: AUC-ROC, sensitivity/specificity of response prediction, top contributing features.
• Difficulty: ⭐⭐⭐⭐ | Novelty: ★★★★★

7. NLP Detection of Depression Symptom Severity from Patient-Generated Text (Social Media / Clinical Notes)

• Research Question: Does NLP analysis of patient-generated text predict PHQ-9 depression severity scores with clinically useful accuracy, and can it identify patients at risk of relapse between clinical visits?
• Why it's novel: Harrison's (2025) explicitly states ML has been applied to "detect depression from patient speech" and "detect depression symptom severity from spoken language." Extending this to written text for outpatient monitoring is a direct PharmD + digital health application.
• Tools you'll use: Python, HuggingFace (BERT/RoBERTa), CLPsych dataset (public NLP mental health dataset), or de-identified clinical notes.
• Outcome: Correlation between NLP-derived severity score and PHQ-9; sensitivity for detecting PHQ-9 ≥10 (moderate-severe depression).
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★

8. Chatbot-Based Medication Adherence Intervention in Major Depressive Disorder: Accuracy and Acceptability

• Research Question: Does an AI chatbot–assisted medication adherence program improve antidepressant adherence (MMAS-8) and depression remission rates (PHQ-9) at 12 weeks in outpatient MDD patients compared to standard care?
• Why it's novel: Goes beyond simple reminders — the chatbot uses conversational AI to provide empathetic, personalized counseling. Kaplan & Sadock highlights digital therapeutics as an emerging research priority.
• Tools you'll use: GPT-4 API or Rasa (open-source chatbot framework), REDCap for data collection, Twilio for SMS delivery. No advanced ML coding needed — primarily a clinical study design.
• Outcome: MMAS-8 score change, PHQ-9 at 12 weeks, chatbot acceptability (SUS scale), safety escalation frequency.
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★

9. Machine Learning–Assisted Pharmacovigilance: Detecting Antidepressant-Induced Suicidality Signals in FAERS

• Research Question: Do ML-based signal detection methods (Random Forest, LASSO regression) identify antidepressant-associated suicidality signals in FAERS that are missed by traditional disproportionality analysis (PRR, ROR)?
• Why it's novel: Combines pharmacovigilance, psychiatry, and AI — a true triple-intersection topic. FDA FAERS is publicly accessible, so no patient recruitment is needed.
• Tools you'll use: FAERS public database (downloadable), Python/R, OpenVigil 2.1 (free pharmacovigilance tool), signal detection algorithms.
• Outcome: New safety signals identified, comparison of ML vs. traditional disproportionality statistics, subgroup risk profiles.
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★

10. Explainable AI (XAI) for Predicting Lithium Toxicity Risk in Bipolar Disorder

• Research Question: Can an XAI model predict lithium toxicity events (serum level >1.5 mEq/L with symptoms) in outpatient bipolar patients with higher sensitivity than standard monitoring protocols alone?
• Why it's novel: Lithium has a narrow therapeutic index — toxicity is common yet often preventable. XAI adds the critical dimension of clinical explainability, making it directly actionable for pharmacists and prescribers.
• Tools you'll use: Python (scikit-learn, SHAP library), hospital EHR data or MIMIC-IV, CPIC lithium dosing guidelines as reference.
• Outcome: Sensitivity/specificity for toxicity prediction, top SHAP-identified risk features, clinical decision support recommendation.
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★

Master Summary Table

Which One Should You Pick?

Sources: Harrison's Principles of Internal Medicine, 22nd Ed. (2025), Chapter 501 — Machine Learning and Augmented Intelligence in Clinical Medicine; Kaplan & Sadock's Comprehensive Textbook of Psychiatry; The Maudsley Prescribing Guidelines, 15th Ed.

PharmD Research Topics Completable in 6 Months

(General Medicine + Psychiatry + AI-Based)

• ✅ No or minimal patient recruitment required
• ✅ Public datasets available OR retrospective chart review
• ✅ Clear, bounded scope
• ✅ No complex lab/biomarker work
• ✅ Publishable output achievable within timeline

🏥 GENERAL MEDICINE

✅ Topic 1 — LLM Accuracy Evaluation for Medication Counseling in Chronic Disease

• Data needed: None — you generate it yourself using LLM APIs (free/cheap)
• Coding needed: None — Excel/SPSS for analysis
• Difficulty: ⭐⭐ | Novelty: ★★★★★
• Why it fits 6 months: Fully self-contained, no IRB patient recruitment, fast data collection

✅ Topic 2 — Pharmacist-Led Deprescribing of Potentially Inappropriate Medications (PIMs) in Elderly Inpatients: A Retrospective Audit

• Data needed: Hospital charts (retrospective — no patient interaction)
• Coding needed: None — Excel/SPSS
• Difficulty: ⭐⭐ | Novelty: ★★★★
• Why it fits 6 months: Retrospective design = no waiting for outcomes

✅ Topic 3 — ML-Based Prediction of 30-Day Readmission in Polypharmacy Patients Using MIMIC-IV

• Data needed: MIMIC-IV (100% free, publicly available at physionet.org)
• Coding needed: Basic Python (many free tutorials available)
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★
• Why it fits 6 months: Public dataset = no IRB patient recruitment delay

🧠 PSYCHIATRY

✅ Topic 4 — Evaluating ChatGPT/LLM Responses for Antidepressant and Antipsychotic Counseling: Safety and Accuracy Study

• Data needed: None — self-generated
• Coding needed: None
• Difficulty: ⭐⭐ | Novelty: ★★★★★
• Why it fits 6 months: Zero recruitment, zero lab work, pure evaluation study

✅ Topic 5 — FAERS Pharmacovigilance: ML-Based Detection of Underreported Adverse Events with Antidepressants or Antipsychotics

• Data needed: FDA FAERS (100% free, downloadable at fda.gov)
• Coding needed: Basic Python or R (or OpenVigil 2.1 — free GUI tool, no coding)
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★
• Why it fits 6 months: Public dataset, no IRB, well-defined analytical pipeline

✅ Topic 6 — Systematic Review + Meta-Analysis: Psilocybin vs. SSRIs in Treatment-Resistant Depression

• Data needed: Published RCTs only — no patient data
• Coding needed: None (RevMan is free and GUI-based)
• Difficulty: ⭐⭐ | Novelty: ★★★★★
• Why it fits 6 months: Systematic reviews are the most time-efficient publishable research format

✅ Topic 7 — NLP Analysis of Patient Reviews of Psychiatric Medications: Extracting Real-World Side Effect Patterns

• Data needed: Publicly available patient reviews (Drugs.com dataset also available on Kaggle — free)
• Coding needed: Basic Python (NLTK / spaCy — many free tutorials)
• Difficulty: ⭐⭐⭐ | Novelty: ★★★★★
• Why it fits 6 months: Dataset ready to use on Day 1, no IRB for public data

Final Comparison Table

🏆 Top 3 Picks for a 6-Month PharmD Project

Disease-Related Observational Studies for PharmD (6 Months)

General Medicine + Psychiatry | Completable with Hospital Access

Observational study types used below:

• Cross-sectional — snapshot at one point in time
• Retrospective cohort — past patient records review
• Prospective cohort — follow patients forward in time
• Case-control — compare cases vs. controls

🏥 GENERAL MEDICINE — Observational Studies

1. Prevalence and Pattern of Comorbidities in Hypertensive Patients on Polypharmacy

• What is the prevalence of comorbid diabetes, CKD, and CVD in hypertensive patients admitted to medicine wards?
• What proportion of patients have ≥1 clinically significant drug-drug interaction?
• Is there a gap between guideline-recommended therapy and actual prescribing?

2. Drug Utilization Study (DUS) of Antidiabetic Agents in Type 2 Diabetes: Adherence to ADA Guidelines

• What is the drug utilization pattern of antidiabetics in T2DM patients in your hospital?
• What percentage of T2DM patients with established CVD are on an SGLT2 inhibitor or GLP-1 agonist per ADA guidelines?
• What clinical and demographic factors are associated with uncontrolled HbA1c (>8%)?

3. Assessment of Medication Adherence and Its Clinical Determinants in Heart Failure Patients

• What is the prevalence of medication non-adherence in HF patients?
• Which medications have the lowest adherence rates (diuretics, beta-blockers, ACE-inhibitors, SGLT2i)?
• What are the predictors of non-adherence (polypharmacy, side effects, cost, health literacy)?

4. Prevalence of Drug-Related Problems (DRPs) in Chronic Kidney Disease (CKD) Patients

• What is the prevalence and type of DRPs in CKD patients in your hospital?
• Which drug classes most commonly require renal dose adjustment but are not adjusted?
• Does CKD stage correlate with increasing number of DRPs?

5. Prescribing Pattern and Outcome of Antibiotics in Community-Acquired Pneumonia (CAP): A Guideline Concordance Study

• What proportion of CAP patients receive guideline-concordant antibiotic therapy?
• Does guideline-concordant prescribing improve clinical outcomes vs. non-concordant prescribing?
• What patient and clinical factors predict guideline deviation?

🧠 PSYCHIATRY — Observational Studies

6. Prevalence and Pattern of Metabolic Syndrome in Patients on Second-Generation Antipsychotics (SGAs)

• What is the prevalence of metabolic syndrome in patients on SGAs?
• Which SGA is most associated with metabolic syndrome?
• Does duration and dose of SGA use predict severity of metabolic abnormalities?

7. Assessment of Medication Adherence and Relapse Rate in Schizophrenia Patients: A 6-Month Prospective Observational Study

• What is the medication adherence rate in outpatient schizophrenia patients?
• What is the 6-month relapse rate and its correlation with adherence?
• Which factors predict non-adherence: side effects, insight, stigma, caregiver support, or polypharmacy?

8. Drug Utilization and Prescribing Pattern of Antidepressants in Major Depressive Disorder: A Retrospective Study

• What is the prescribing pattern of antidepressants in your hospital?
• What proportion of patients receive guideline-recommended first-line therapy?
• What is the rate of antidepressant switching, augmentation, and polypharmacy?

9. Prevalence of Polypharmacy and Drug-Drug Interactions in Hospitalized Psychiatric Patients

• What is the prevalence of polypharmacy (≥5 drugs) in psychiatric inpatients?
• What is the frequency and severity distribution of drug-drug interactions?
• Which drug combinations are most commonly involved in major DDIs?

10. Quality of Life and Functional Outcomes in Bipolar Disorder Patients on Mood Stabilizers: A Cross-Sectional Study

• What is the quality of life in bipolar patients during euthymia?
• Does type of mood stabilizer correlate with QoL and functional outcomes?
• Is medication adherence a significant predictor of QoL in bipolar disorder?

Master Summary Table

🏆 Top Picks by Goal

6-Month Generic Timeline (Applies to All Topics)

Research Gap Analysis

Article: "Prevalence of Polypharmacy and Associated Adverse Health Outcomes in Adult Patients with Chronic Kidney Disease: Protocol for a Systematic Review and Meta-Analysis"

What This Article Covers (Scope of the Study)

• Prevalence of polypharmacy (≥5 drugs) in adult CKD patients (all stages including dialysis)
• Association between polypharmacy and adverse health outcomes (AKI, CVD events, mortality, hospitalizations, ADRs, HRQoL)
• Subgroup analysis by sex, age, KRT status, and multimorbidity
• Study population: Adults ≥18 years, inpatient and outpatient settings

Identified Research Gaps

GAP 1 — No Focus on Pediatric or Adolescent CKD Patients

"Prevalence and pattern of polypharmacy and drug-related problems in pediatric CKD patients: a cross-sectional observational study"

• Why novel: Findings from adult CKD cannot be extrapolated to children due to developmental pharmacokinetics and weight-based dosing complexities.

GAP 2 — No Assessment of Pharmacist Interventions or Deprescribing Strategies

"Impact of pharmacist-led medication review and deprescribing on polypharmacy-related outcomes in CKD patients: a prospective observational study"

• Why novel: Directly addresses the "key priority area to improve medication practices" the authors themselves call for in the Discussion section (p.5).

GAP 3 — No Inclusion of Over-the-Counter (OTC) Drugs, Herbal, and Complementary Medicines

"Prevalence and nephrotoxic risk of over-the-counter medications, herbal products, and dietary supplements used concurrently with prescription drugs in CKD patients: a cross-sectional observational study"

• Why novel: True polypharmacy burden in CKD is underestimated when OTC/herbal use is excluded — a critical PharmD patient safety research area.

GAP 4 — No Disease-Specific or CKD-Stage–Specific Drug Utilization Analysis

"Drug utilization pattern and prevalence of renally inappropriate medications (RIMs) across CKD stages G3–G5D: a cross-sectional observational study using KDIGO staging"

• Why novel: Identifies the specific CKD stage at which polypharmacy risk peaks and which drug classes need targeted intervention — actionable data for nephrology pharmacists.

GAP 5 — No Assessment of Quality of Life (QoL) as a Primary Outcome

"Impact of polypharmacy on health-related quality of life (HRQoL) in CKD patients: a prospective cross-sectional observational study using KDQOL-36 and EQ-5D"

• Why novel: Shifts the focus from clinical events to patient-reported outcomes — aligned with modern patient-centered care models and PCORI research priorities.

GAP 6 — No Evaluation of Medication Non-Adherence as an Outcome

"Association between polypharmacy and medication non-adherence in CKD patients: a prospective observational study using MMAS-8 and pill count"

• Why novel: Directly fills the gap noted in Table 1 of the article itself. Connects polypharmacy burden directly to adherence behavior — a measurable, modifiable PharmD-relevant outcome.

GAP 7 — No Low- and Middle-Income Country (LMIC) Data

"Prevalence of polypharmacy, potentially inappropriate medications, and renally inappropriate prescribing in CKD patients in a tertiary care hospital in [your country]: a cross-sectional observational study"

• Why novel: Provides local/regional epidemiological data that the global systematic review cannot capture — highly publishable in regional nephrology and pharmacy journals.

GAP 8 — No Role of the Pharmacist or Multidisciplinary Team Assessed

"Role of the clinical pharmacist in identifying and resolving drug-related problems in hospitalized CKD patients: a prospective interventional observational study"

• Why novel: Converts the review's descriptive findings into actionable PharmD practice research.

GAP 9 — No AI or Predictive Modeling Approach for Early Polypharmacy Detection

"Development and validation of a machine learning model to predict high-risk polypharmacy and ADEs in CKD patients using EHR data"

• Why novel: The article's own risk factors (CKD stage, age, BMI, comorbidities) are ideal ML features — this is a direct, evidence-based AI extension of the systematic review.

Summary Table of All Research Gaps

Best Research Topics Derived From These Gaps (Ready to Use)

Research Proposal

"Drug Utilization Pattern and Prevalence of Renally Inappropriate Medications (RIMs) Across CKD Stages G3–G5D: A Cross-Sectional Observational Study Using KDIGO Staging"

1. AIM

2. OBJECTIVES

Primary Objective

• To determine the prevalence of renally inappropriate medications (RIMs) — defined as drugs that are contraindicated, require dose reduction, or require dose interval adjustment based on the patient's eGFR — across CKD stages G3, G4, G5, and G5D.

Secondary Objectives

1. To describe the drug utilization pattern (drug classes, average number of drugs, polypharmacy rate) across each CKD stage.
2. To identify the most commonly prescribed RIMs and the drug classes most frequently involved.
3. To determine the prevalence of polypharmacy (≥5 drugs) and hyperpolypharmacy (≥10 drugs) across CKD stages.
4. To assess the proportion of patients receiving drugs contraindicated in CKD vs. those receiving drugs that require dose adjustment but were not adjusted.
5. To evaluate the association between CKD stage severity and the number of RIMs per patient.
6. To identify patient-level risk factors (age, sex, comorbidities, number of prescribers) associated with RIM exposure.

3. INCLUSION CRITERIA

4. EXCLUSION CRITERIA

5. SAMPLE SIZE

Calculation Basis

• Z = 1.96 (95% confidence interval)
• P = Expected prevalence of RIMs in CKD = 0.62 (62%), based on published data showing PIMs/RIMs prevalence of 62–67% in CKD inpatient and ambulatory settings (cited in Okpechi et al., 2021, referencing Blix et al., 2006)
• d = Margin of error = 0.07 (7%)

Distribution Across CKD Stages

Sampling method: Consecutive sampling of eligible patients from nephrology OPD and wards over the study period until target per stage is reached.

6. METHODOLOGY

6.1 Study Design

6.2 Study Setting

• Nephrology OPD and Nephrology/Internal Medicine wards of a tertiary care hospital
• Duration: 6 months (Months 2–5 for data collection, after ethics approval in Month 1)

6.3 Study Period & Timeline

6.4 Data Collection Procedure

• Screen all patients admitted to nephrology ward or attending nephrology OPD daily
• Verify inclusion/exclusion criteria
• Obtain written informed consent

• Calculate eGFR using CKD-EPI 2021 creatinine equation
• Classify albuminuria: A1 (<30 mg/g), A2 (30–300 mg/g), A3 (>300 mg/g)
• Assign CKD stage: G3a, G3b, G4, G5, or G5D
• Document dialysis modality for G5D patients (haemodialysis vs. peritoneal dialysis)

• Prescription charts / inpatient drug charts
• Outpatient prescription records
• Patient-reported medication list (cross-checked)
• Include: drug name (generic), dose, frequency, route, duration of use
• Classify each drug: cardiovascular, antidiabetic, analgesic, antibiotic, GI, psychiatric, haematological, mineral/bone disorder, etc.

• Calculate Defined Daily Dose (DDD) per 100 bed-days (for inpatients)
• Calculate Polypharmacy rate (≥5 drugs) and Hyperpolypharmacy rate (≥10 drugs)
• Calculate Average number of drugs per patient per CKD stage

• Age, sex, weight, BMI
• Primary CKD etiology (diabetic nephropathy, hypertensive nephropathy, glomerulonephritis, unknown)
• Comorbidities (diabetes, hypertension, heart failure, dyslipidemia, gout, anemia)
• Number of prescribing physicians involved in care
• Duration of CKD diagnosis
• Serum creatinine, eGFR, albumin, electrolytes (K⁺, Na⁺, PO₄, Ca²⁺), CBC

6.5 Data Recording Tool

1. Patient demographics and CKD staging
2. Complete medication list with dose and frequency
3. RIM classification checklist (Type 1/2/3 per drug)

6.6 Statistical Analysis

7. OUTCOME MEASURES

Primary Outcome

Secondary Outcomes

8. ETHICAL CONSIDERATIONS

• Ethics/IRB approval required prior to data collection
• Written informed consent from all patients (or legal representative)
• Patient anonymity maintained — data recorded with unique ID codes only
• No intervention — purely observational; patient management will not be altered by the study
• Data security — password-protected spreadsheet; accessible only to research team

9. EXPECTED OUTCOMES & SIGNIFICANCE

10. LIMITATIONS (Acknowledged)

References for methodology: Brenner & Rector's The Kidney (KDIGO CGA staging, p. 841); Okpechi et al., Systematic Reviews (2021); KDIGO CKD Clinical Practice Guidelines (2012, 2024 update); Aronoff et al., Drug Prescribing in Renal Failure, 5th ed.; WHO Drug Utilization Methodology (DDD framework).

Revised Sample Size Section

5. SAMPLE SIZE

Approach: Adjusted Parameters for Feasibility-Based Sample Size

Formula Used

✅ Final sample size = 99 patients — within the <100 target and statistically justified at 95% CI with 10% margin of error.

Why a 10% Margin of Error is Acceptable Here

Distribution Across CKD Stages

Sampling method: Consecutive sampling — all eligible patients attending nephrology OPD or admitted to nephrology ward are enrolled in sequence until each stage quota is met.

Enrollment Rate Feasibility Check

Summary: A sample of 99 patients is statistically justified at 95% confidence level with a 10% margin of error, epidemiologically appropriate for a single-centre pilot cross-sectional study, and practically achievable within a 6-month PharmD project timeline.

Research Guide's Recommendation

Observational Research Topics — 6-Month PharmD Project

General Medicine + Psychiatry Department

As your research guide, I have applied these filters before recommending any topic:

• ✅ Completable by one PharmD student in 6 months
• ✅ Requires no advanced lab work or expensive investigations
• ✅ Needs sample size ≤100 — realistic for a single centre
• ✅ Has a clear, measurable primary outcome
• ✅ Has a research gap you can justify from published literature
• ✅ Publishable in a pharmacy practice or clinical medicine journal
• ✅ Directly demonstrates the value of a pharmacist in clinical care

🏥 GENERAL MEDICINE DEPARTMENT

⭐ GUIDE'S TOP PICK — General Medicine

Topic GM-1

"Prevalence of Drug-Related Problems (DRPs) and Associated Risk Factors in Hospitalized Patients with Type 2 Diabetes Mellitus: A Prospective Cross-Sectional Observational Study"

Type 2 diabetes patients are on multiple drug classes simultaneously — antidiabetics, antihypertensives, statins, antiplatelet drugs. This creates a high DRP burden that is directly identifiable, classifiable, and reportable by a pharmacist. The topic is clinically relevant, has a well-established methodology, and produces data that can directly improve prescribing practices in your hospital.

• Formula: n = Z² × P × (1−P) / d²
• P = 0.65 (65% DRP prevalence in diabetes inpatients — published data)
• Z = 1.96, d = 0.10
• n = 3.84 × 0.65 × 0.35 / 0.01 = 87 patients
• +10% attrition = 96 patients ✅ (under 100)

1. To determine the prevalence and types of DRPs in hospitalized T2DM patients using PCNE v9.1
2. To identify the drug classes most frequently involved in DRPs
3. To assess the association between polypharmacy and number of DRPs
4. To identify patient-level predictors of DRPs (age, comorbidity count, HbA1c, number of prescribers)

Topic GM-2

"Assessment of Prescribing Patterns and Guideline Concordance of Antihypertensive Therapy in Patients with Hypertension and Comorbid Type 2 Diabetes: A Retrospective Drug Utilization Study"

Hypertension + diabetes is the most common comorbid combination in general medicine wards worldwide. ACC/AHA 2017 and JNC 8 guidelines clearly define first-line therapy (ACE inhibitor or ARB mandatory for diabetic hypertensives). Despite this, real-world prescribing deviations are common and under-documented locally. This retrospective DUS requires only prescription records — zero patient recruitment burden.

• P = 0.55 (55% guideline non-concordance — conservative estimate from published DUS literature)
• Z = 1.96, d = 0.10
• n = 3.84 × 0.55 × 0.45 / 0.01 = 95 patients ✅

1. To describe the prescribing pattern of antihypertensive drugs in patients with hypertension + T2DM
2. To assess guideline concordance with ACC/AHA 2017 / JNC 8 recommendations
3. To calculate DDD (Defined Daily Dose per 100 OPD visits) for each antihypertensive class
4. To determine the proportion of diabetic hypertensives receiving ACE inhibitor or ARB as mandated by guidelines
5. To identify factors associated with guideline-discordant prescribing

Topic GM-3

"Prevalence of Potentially Inappropriate Medications (PIMs) in Elderly Patients (≥60 years) Admitted to General Medicine Ward: A Cross-Sectional Study Using Beers Criteria 2023"

Elderly polypharmacy is a global crisis. The 2023 AGS Beers Criteria is the most recent update (released 2023) — very few Indian/Asian hospital studies have used the 2023 version yet. This creates an immediate, publishable gap. It is a straightforward checklist-based observational study — you compare the patient's drug list against the Beers Criteria. No complex investigations needed.

🧠 PSYCHIATRY DEPARTMENT

⭐ GUIDE'S TOP PICK — Psychiatry

Topic PSY-1

"Prevalence of Metabolic Syndrome and Its Association with Second-Generation Antipsychotic Use in Patients with Schizophrenia: A Cross-Sectional Observational Study"

This is one of the most clinically important and under-addressed problems in psychiatric pharmacotherapy. Second-generation antipsychotics (SGAs) — olanzapine, clozapine, quetiapine, risperidone — are associated with significant weight gain, dyslipidemia, and glucose dysregulation, collectively causing metabolic syndrome. This is directly confirmed in The Maudsley Prescribing Guidelines (15th ed.) which cites CATIE trial data showing SGA-induced metabolic parameter changes. Metabolic syndrome leads to cardiovascular mortality — the #1 cause of premature death in schizophrenia patients. As a pharmacist, identifying this is your direct clinical role.

• P = 0.40 (40% metabolic syndrome prevalence in SGA-users — conservative; published range 30–60%)
• Z = 1.96, d = 0.10
• n = 3.84 × 0.40 × 0.60 / 0.01 = 92 patients
• +5% attrition = 97 patients ✅ (under 100)

1. To determine the prevalence of metabolic syndrome in schizophrenia patients receiving SGAs using IDF criteria
2. To compare metabolic syndrome prevalence across SGA agents (olanzapine, risperidone, quetiapine, clozapine, aripiprazole)
3. To assess the association between SGA dose, duration of use, and severity of metabolic abnormalities
4. To identify patient-level risk factors for SGA-induced metabolic syndrome (age, sex, BMI at baseline, family history of diabetes/CVD)
5. To determine the proportion of patients not being monitored for metabolic parameters per guideline recommendations

Topic PSY-2

"Assessment of Medication Non-Adherence and Its Clinical Predictors in Outpatients with Major Depressive Disorder on Antidepressant Therapy: A Cross-Sectional Observational Study"

Non-adherence to antidepressants is estimated at 40–60% in outpatient MDD patients. It is the single biggest cause of treatment failure, relapse, and recurrence. As a pharmacist, this is your domain — identifying why patients don't take their medications and what predicts non-adherence. This study is simple, clean, and publishable within 6 months using only a validated questionnaire and medication records.

1. To assess the prevalence of medication non-adherence in MDD outpatients using MMAS-8 scale
2. To identify patient-level predictors of non-adherence (side effects, cost, stigma, polypharmacy, insight)
3. To correlate adherence score with depression severity (PHQ-9)
4. To determine which antidepressant classes have the lowest adherence rates

Topic PSY-3

"Drug-Drug Interaction Profile in Patients Receiving Antipsychotic Polypharmacy: A Cross-Sectional Observational Study in a Tertiary Psychiatry Unit"

Psychiatric patients are frequently on 2 or more antipsychotics (antipsychotic polypharmacy — APP) plus adjunct drugs (mood stabilizers, anticholinergics, benzodiazepines, antidepressants). DDIs in this setting are common, often clinically significant (QTc prolongation, EPS, sedation, metabolic worsening), and directly preventable by a pharmacist. This is a checklist-based observational study — you simply document all drugs and run them through Lexicomp/Micromedex.

GUIDE'S FINAL RECOMMENDATION TABLE

AS YOUR GUIDE — My Final Advice

If you want one topic from each department, I would recommend:

Both studies can be run simultaneously in the same 6-month period since they are in different departments — making your project broader, stronger, and more publishable as a comparative or two-part study.

Research Guide's Fresh Topic List

Observational Studies | General Medicine + Psychiatry

Strictly Excluding: Polypharmacy · DDIs · PIMs/RIMs

🏥 GENERAL MEDICINE — Fresh Topics

✅ GM-1 — GUIDE'S TOP PICK

"Assessment of Knowledge, Attitude, and Practice (KAP) Regarding Disease Self-Management Among Patients with Type 2 Diabetes Mellitus: A Cross-Sectional Observational Study"

1. To assess the level of knowledge about T2DM self-management (diet, exercise, blood glucose monitoring, foot care, medication)
2. To evaluate patient attitudes toward disease management and lifestyle modification
3. To assess current self-care practices (SDSCA — Summary of Diabetes Self-Care Activities scale)
4. To identify predictors of poor KAP (education level, duration of disease, income, number of OPD visits)
5. To correlate KAP score with glycemic control (HbA1c)

✅ GM-2

"Assessment of Cardiovascular Risk Factors and 10-Year CVD Risk Using Framingham Risk Score in Patients with Hypertension Attending a Tertiary Care OPD: A Cross-Sectional Study"

1. To determine the prevalence of modifiable cardiovascular risk factors (dyslipidemia, diabetes, obesity, smoking, physical inactivity) in hypertensive patients
2. To calculate 10-year CVD risk using the Framingham Risk Score
3. To categorize patients into low (<10%), intermediate (10–20%), and high (>20%) CVD risk groups
4. To assess whether antihypertensive treatment intensity (number of agents, BP target achievement) is appropriately aligned with CVD risk category
5. To identify the most prevalent modifiable risk factor amenable to pharmacist-led intervention

✅ GM-3

"Assessment of Nutritional Status and Its Association with Clinical Outcomes in Hospitalized Patients with Chronic Liver Disease: A Cross-Sectional Observational Study"

1. To determine the prevalence of malnutrition/malnutrition risk in hospitalized CLD patients using RFH-NPT
2. To assess the association between nutritional status and disease severity (Child-Pugh class A/B/C)
3. To correlate nutritional status with length of hospital stay, ascites severity, hepatic encephalopathy episodes, and 30-day readmission
4. To identify clinical and laboratory predictors of malnutrition (serum albumin, BMI, mid-arm circumference)

✅ GM-4

"Evaluation of Blood Pressure Control and Associated Factors in Hypertensive Patients on Antihypertensive Therapy: A Cross-Sectional Observational Study"

1. To determine the proportion of hypertensive patients achieving target BP (<130/80 mmHg)
2. To assess medication adherence using MMAS-8 and its correlation with BP control
3. To identify patient-level factors associated with uncontrolled BP: age, sex, duration of hypertension, number of antihypertensives, comorbidities, lifestyle factors
4. To assess the association between number of antihypertensive agents and BP control rates

🧠 PSYCHIATRY DEPARTMENT — Fresh Topics

✅ PSY-1 — GUIDE'S TOP PICK

"Assessment of Adverse Drug Reaction Profile of Second-Generation Antipsychotics and Its Impact on Medication Adherence in Schizophrenia Patients: A Cross-Sectional Observational Study"

• UKU Side Effect Rating Scale (gold standard for antipsychotic ADR assessment — covers 48 items across 4 domains: psychic, neurologic, autonomic, other)
• MMAS-8 for medication adherence
• PANSS (optional, for symptom severity correlation)

1. To assess the type, frequency, and severity of ADRs in schizophrenia patients receiving SGAs using UKU scale
2. To compare ADR profiles across individual SGA agents (olanzapine vs. risperidone vs. quetiapine vs. aripiprazole vs. clozapine)
3. To determine the association between ADR burden (UKU score) and medication adherence (MMAS-8)
4. To identify which ADR domain (psychic / neurologic / autonomic / other) most strongly predicts non-adherence
5. To assess whether patients are being monitored for specific ADRs per Maudsley/APA guidelines

✅ PSY-2

"Assessment of Caregiver Burden and Its Determinants in Family Members of Patients with Schizophrenia: A Cross-Sectional Observational Study"

1. To assess the level of caregiver burden using ZBI-22 in caregivers of schizophrenia patients
2. To identify patient-level factors associated with higher caregiver burden (symptom severity, duration of illness, number of hospitalizations, ADR burden)
3. To examine caregiver-level predictors (age, sex, relationship to patient, employment status, hours of caregiving per day)
4. To assess whether patient's medication adherence is associated with caregiver burden score

✅ PSY-3

"Assessment of Sleep Quality and Its Association with Depression and Anxiety in Patients with Major Depressive Disorder: A Cross-Sectional Observational Study"

• PSQI (Pittsburgh Sleep Quality Index) — sleep quality
• PHQ-9 — depression severity
• GAD-7 — anxiety severity

1. To assess the prevalence of poor sleep quality (PSQI >5) in MDD patients on antidepressant therapy
2. To determine the correlation between PSQI score and PHQ-9 depression severity
3. To assess the association between sleep quality and anxiety levels (GAD-7)
4. To identify predictors of poor sleep in MDD patients (age, sex, antidepressant type, duration of illness, comorbid anxiety)
5. To determine whether patients with poor sleep quality have worse antidepressant treatment response

✅ PSY-4

"Assessment of Metabolic Syndrome Monitoring Practices and the Gap Between Guideline Recommendations and Real-World Monitoring in Patients on Long-Term Antipsychotic Therapy: A Cross-Sectional Observational Study"

1. To assess the proportion of patients on long-term antipsychotic therapy receiving guideline-recommended metabolic monitoring (weight, BMI, fasting glucose, lipids, BP)
2. To compare monitoring rates across antipsychotic agents (clozapine — highest risk — vs. others)
3. To identify factors associated with monitoring gaps (outpatient vs. inpatient, prescriber type, disease duration)
4. To document the prevalence of undetected metabolic abnormalities in patients with no recent monitoring

FINAL GUIDE'S RECOMMENDATION TABLE

My Single Best Pair for Your Project