NR 538 Week 2 Data Source Appraisal Example

Data Source Appraisal: The County Ambulance Service's Patient Care Records, Built to Bill and to Report Responses, Tested Field by Field for Which of Their Jobs Also Serve a Falls Assessment and Which Do Not

Student Name

College of Nursing, Chamberlain University

NR 538: Population Health I: Assessment and Analysis of Data

Instructor Name

Month Day, Year

Data Source Appraisal: The County Ambulance Service's Patient Care Records, Built to Bill and to Report Responses, Tested Field by Field for Which of Their Jobs Also Serve a Falls Assessment and Which Do Not

The Source and Its Owner's Purpose

This appraisal examines the electronic patient care records of our county's ambulance service as a source for assessing older adults at risk of serious falls. Somebody paid for every row, and it was not me. The service keeps these records for three reasons: to bill Medicare, Medicaid and insurers for transports; to document care for clinical and legal purposes; and to submit a standard data set to the state emergency medical services office, which forwards it to the national EMS information system. A description of the national data set shows that it collects millions of activations each year from thousands of agencies, with standard fields for dispatch reason, symptoms, procedures and disposition (Mann et al., 2015).

Each of those purposes makes some fields reliable and leaves others unchecked. The appraisal matters because the assessment team is about to use the file for a purpose its owner never had in mind: finding people at risk, rather than documenting a response. A file can be excellent for its owner and misleading for a secondary user, and the difference usually lies in fields the owner never had reason to check. The same records, for instance, would serve well for studying response times, because every timestamp is automatic, and poorly for studying the circumstances of falls, because nothing requires crews to describe them. Fields that affect payment or legal defense are completed carefully and audited. Fields that no one downstream uses may be left blank, filled with defaults or entered inconsistently. The appraisal therefore asks, field by field, whether the owner's purpose keeps it accurate, and whether that accuracy extends to the use I intend: identifying and describing older adults who fall and cannot get up.

Fields That Serve the Assessment Well

Date, time and address of the call. These are recorded automatically by the dispatch system and are needed for billing and for the state report. They are reliable. The address allows repeat calls to the same home to be linked, and geocoding allows calls to be mapped.

Age and sex. Recorded for every patient contact and required by the state data set. Reliable, with occasional estimated ages when a patient cannot give a date of birth.

Disposition. Whether the patient was transported, refused transport or was helped without transport is recorded carefully, because it determines whether a bill is generated. This is the field that identifies lift-assist calls. It is reliable for my purpose.

Dispatch reason. The dispatcher's code for the call, such as "fall" or "lift assist," is recorded consistently because the dispatch protocol requires it. It is less precise than the crew's findings but reliable as a first filter.

Fields That Do Not

Cause of the fall and circumstances. The records contain a free-text narrative in which crews may describe how the person fell. Nobody downstream uses it for billing or reporting, and completeness varies by crew. In a sample of 40 lift-assist records from 2025, 14 narratives described the circumstances, such as tripping on a rug or getting up at night, and 26 said only "patient on floor, assisted to chair." The narrative cannot be used to estimate causes.

Injury assessment. For patients not transported, crews complete a brief assessment, but documentation of injury is inconsistent, because there is no bill to support and no required field for minor injuries. The file cannot reliably show how many lift-assist patients had minor injuries.

Linking the same person across calls. The records use name and date of birth, which are sometimes misspelled or missing for patients who are confused. Address linkage is more reliable than name linkage, but it will occasionally merge two older adults in the same household.

Coverage and Completeness

The file covers only falls that reach the ambulance service. It misses falls handled by family, falls in the two assisted living facilities whose staff rarely call for lifts, and falls whose victims are driven to the emergency department by relatives. It therefore undercounts falls and overrepresents older adults who live alone or with a frail spouse. This is a bias I can describe but not correct from this source. A study of lift-assist calls in one American city found that two-thirds of calls went to one-third of the addresses, with many repeat visits within 30 days (Cone et al., 2013). If our pattern is similar, the file will be most useful for identifying repeat callers, who are the people most at risk.

Timeliness, Access and Ethics

The file's history also matters. The service adopted its current record system four years ago, and records from before that date use a different set of disposition codes, so trend analysis across the change will need care. I asked the service's data manager what changed and received a one-page crosswalk of old and new codes, which will be essential when earlier years are compared. Records are finalized within 24 hours of a call and are available to the county through a data use agreement renewed annually. The agreement allows de-identified extracts for public health purposes and identified data only for approved outreach programs, with the ambulance service's medical director's consent. Any outreach to identified lift-assist callers would need a separate agreement and a plan for how people are contacted, since some may not expect a health department to know they called an ambulance.

Verdict

The file is suitable for three uses in this assessment: counting lift-assist calls and repeat callers among older adults, describing them by age, sex, time and location, and mapping them across the county. Those three uses cover most of what the assessment needs from this source, and they rest on the fields the owner keeps most carefully. It is not suitable for estimating the causes of falls, the frequency of minor injuries or the total number of falls among older adults. For those, I will need other sources: hospital discharge data for injuries serious enough to require admission, the state's survey estimate of self-reported falls for the total and, if possible, a small sample of home visits to lift-assist callers for causes. The appraisal's main lesson is that a file built to bill and report can answer questions that follow its owner's purpose and cannot answer questions its owner never needed to ask.

References

Cone, D. C., Ahern, J., Lee, C. H., Baker, D., Murphy, T., & Bogucki, S. (2013). A descriptive study of the "lift-assist" call. Prehospital Emergency Care, 17(1), 51-56. https://doi.org/10.3109/10903127.2012.717168

Mann, N. C., Kane, L., Dai, M., & Jacobson, K. (2015). Description of the 2012 NEMSIS public-release research dataset. Prehospital Emergency Care, 19(2), 232-240. https://doi.org/10.3109/10903127.2014.959219