| Course | HLTH 9101 Dissertation |
|---|---|
| Week | Week 5 |
| Paper type | Method write-up |
| Length | About 600 words, 5 pages |
| Format | APA 7 student paper |
| School | Walden University |
| Program | PhD in Public Health |
| Updated | October 2026 |
Free sample paper for HLTH 9101 Week 5
Method Write-Up for a Stranger: How a Composite County-Level Study Went From Public Downloads to Models, Which Counties Dropped Out and Why, and the Diagnosis, Procedure and Modifier Codes That Turned Claims Into Behavioral Telehealth Visits
Student Name
Walden University
HLTH 9101: Dissertation
Instructor Name
Month Day, Year
Method Write-Up for a Stranger: How a Composite County-Level Study Went From Public Downloads to Models, Which Counties Dropped Out and Why, and the Diagnosis, Procedure and Modifier Codes That Turned Claims Into Behavioral Telehealth Visits
Standard for This Chapter
This method chapter is written for an analyst who has never met the author and wants to rebuild the study from public files. Every step from download to model is described in order, sources are named by type, and every exclusion is counted. The reporting follows the RECORD extension of STROBE for studies using routinely collected health data, which asks authors to describe how codes were used to define study variables (Benchimol et al., 2015).
Design and Units
The study is a cross-sectional ecological analysis with the county as the unit of analysis, covering the 2022 calendar year. The population is Medicare fee-for-service beneficiaries aged 65 and older with at least one behavioral health visit in 2022, aggregated to their county of residence.
Sources
Three kinds of public files were used. The exposure, the percentage of households with a broadband internet subscription, came from a national household survey's five-year county estimates ending in 2022. The outcome came from an aggregated Medicare claims file that reports, by county, counts of behavioral health visits and visits billed as telehealth for beneficiaries aged 65 and older. Covariates came from the same household survey (median household income, share aged 65 and older), a federal rural classification file, a national clinician registry (behavioral clinicians for every ten thousand people) and a Medicare enrollment file (share of beneficiaries in Medicare Advantage).
Defining a Behavioral Health Telehealth Visit
The primary outcome is telehealth visits divided by all behavioral health visits, expressed as a percentage. The secondary outcome is video visits, telehealth visits without an audio-only modifier, divided by all behavioral health visits.
Table 1
Codes Used to Define the Outcome
| Element | Codes | Purpose | |
|---|---|---|
| Behavioral health diagnosis | ICD-10-CM F01-F99 as primary diagnosis | Restricts visits to mental, behavioral and substance use conditions | |
| Behavioral health services | CPT 90791-90792 (diagnostic evaluation), 90832-90838 (psychotherapy), 99212-99215 with an F-code primary diagnosis | Captures assessment, therapy and medication management visits | |
| Telehealth, any modality | Place of service 02 or 10, or modifier 95 or GT | Marks visits delivered remotely | |
| Audio-only | Modifier 93 or FQ | Separates telephone visits for the secondary outcome | |
Codes reflect billing rules in effect for 2022; the aggregated file applied these definitions before release.
Counties Included and Excluded
Of 3,143 counties and county equivalents, 2,961 entered the analysis. The aggregated claims file suppresses counts below 11 to protect privacy, which removed 164 counties with too few behavioral health visits; most were rural and small. Eighteen counties lacked a covariate, usually clinician supply. The excluded counties had lower broadband subscription on average, so the analysis underrepresents the least connected places. The counts are illustrative for this composite study.
Analysis
Each county's outcome was modeled with weighted linear regression, weighting by the number of behavioral health visits so that estimates reflect where care occurred. The model included broadband subscription, centered and scaled per 10 percentage points, and all covariates, with standard errors clustered by state. A sensitivity analysis used a fractional logit model suited to proportions. The secondary outcome was modeled the same way. In line with the behavioral model of Andersen (1995), covariates were grouped into predisposing, enabling and need factors in the presentation of results.
What a Stranger Should Know
Results describe counties, not people. Robinson (1950) showed that correlations across groups can differ from those among individuals, so no individual-level conclusion follows. The outcome counts only beneficiaries who received behavioral health care.
References
Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36(1), 1-10. https://doi.org/10.2307/2137284
Benchimol, E. I., Smeeth, L., Guttmann, A., Harron, K., Moher, D., Petersen, I., Sørensen, H. T., von Elm, E., & Langan, S. M. (2015). The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLOS Medicine, 12(10), Article e1001885. https://doi.org/10.1371/journal.pmed.1001885
Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15(3), 351-357. https://doi.org/10.2307/2087176
HLTH 9101 Week 5 instructions, in plain terms
Students write the method chapter of their dissertation so that another analyst could reproduce the study. Faculty expect a clear statement of design, unit of analysis and time period; sources named by type with enough detail to locate them; every variable defined, including the codes used to build outcomes from administrative data; every exclusion counted with reasons and with a note on how excluded units differ; and the analytic model described fully, including weighting, standard errors and sensitivity analyses. Reporting guidelines for routinely collected data, such as RECORD, provide a useful checklist. A table of codes is often essential. The chapter should also state what the design cannot support, so readers know the limits before they see results.
Inside the HLTH 9101 Week 5 example
The chapter opens with its standard and cites Benchimol and colleagues on the RECORD guideline. A short section states the design, unit and year. The sources section describes the household survey, the aggregated claims file and four covariate sources. Table 1 lists diagnosis codes, procedure codes, telehealth place-of-service codes and modifiers, and audio-only modifiers. A paragraph defines the primary and secondary outcomes. The inclusion section counts counties excluded for suppression and missing data and notes that they were less connected. The analysis section describes weighted regression, state-clustered errors, a fractional logit sensitivity model and grouping covariates by Andersen's model. A closing section cites Robinson on ecological inference.
HLTH 9101 Week 5 rubric: what earns full marks
Reproducibility is the standard here, and graders test it by asking whether a stranger could rebuild the analytic file. A method chapter that says outcomes were derived from claims, without codes, fails that test. Strong chapters list every code, count every exclusion and describe how excluded units differ from included ones. The model description should include weighting, the variance approach and sensitivity analyses. Citing a reporting guideline shows awareness of field standards. Graders also value a plain statement of the level at which results can be interpreted. A well-organized table of codes and a clear exclusion count are often the elements that separate strong method chapters from adequate ones.
Common HLTH 9101 Week 5 mistakes, and how to avoid them
Method chapters often describe data sources and analyses in general terms that make sense to the author but not to anyone else. To write a stronger chapter, imagine an analyst with only your public files. Write down every step they would need: where each file comes from, how variables were built, which codes define the outcome, which units were dropped and why, and the exact model. Put codes in a table. Count exclusions and say how the excluded differ. Use a reporting checklist to catch gaps. If your study uses different data, a method chapter built to this standard can be drafted. Then hand the chapter to someone outside your project and ask where they would get stuck.
Get a HLTH 9101 Week 5 example written to your instructions
Provide your data sources, the steps that shaped your analytic sample, variable definitions and planned models, along with the Week 5 prompt and rubric. The chapter is written so an outside analyst could rebuild your work from it, with linkage, exclusions and code lists stated. First write-up free, back in 24 to 48 hours. Every file in the example is generic.
More HLTH 9101 papers and PhD in Public Health samples
- HLTH 9101 Week 1: Dissertation Progress Post
- HLTH 9101 Week 2: Dissertation Argument Audit
- HLTH 9101 Week 3: Literature Synthesis Section
- HLTH 9101 Week 4: Literature Review Chapter Draft
- HLTH 9101 Week 6: Key Term Audit
- HLTH 9101 Week 7: County Findings Chapter
- HLTH 9101 Week 8: Dissertation Discussion Draft
- HLTH 9101 Week 9: Manuscript Consistency Peer Replies
- HLTH 9101 Week 10: Revision by Subtraction
- HLTH 9101 Week 11: Full Dissertation Read
- HLTH 8815 Week 5: Care Setting Comparison
- HLTH 8745 Week 5: Cortisol Measurement Critique
- HLTH 8701 Week 5: Measurement Invariance Appraisal
- HLTH 8574 Week 5: Sjögren's Condition Brief
HLTH 9101 Week 5 questions, answered
How much detail does a method write-up need?
Enough that someone with the same public files could reach the same analytic sample and run the same model. That usually means naming each source's release, describing every join and exclusion, and listing the codes behind each variable. The example puts code lists in an appendix to keep the prose readable. Details you would need to remember a year from now are details a stranger needs today.
What is RECORD and when does it apply?
It is a reporting guideline extending STROBE for studies built on data collected for other purposes, such as claims, registries and administrative files. It asks, among other things, how databases were linked and how variables were identified in the data. The example uses it because its outcome comes from claims. If your data were gathered for your study, STROBE alone may suffice; your committee may state a preference.
Should sensitivity analyses be decided in advance?
Ideally, yes, and the chapter should say which were. A sensitivity analysis chosen after seeing results can look like searching for a preferred answer. The example names two in advance, one excluding audio-only visits and one limited to counties with larger visit totals, and would label any later additions exploratory. If you added one late, say so; readers accept honest labels readily.
Where can I find a free HLTH 9101 Week 5 sample paper?
The full Week 5 method write-up appears here, describing a composite county-level study step by step with its codes, exclusions and models so a stranger could reproduce it, with three references.
Which codes define a behavioral health telehealth visit?
A primary ICD-10-CM diagnosis in F01-F99, a psychotherapy, diagnostic evaluation or evaluation and management procedure code, and a telehealth place-of-service code or modifier, with audio-only modifiers used to separate telephone visits.